diff --git a/.agents/skills/augustus/SKILL.md b/.agents/skills/augustus/SKILL.md index 5313f07a..dbef4c3d 100644 --- a/.agents/skills/augustus/SKILL.md +++ b/.agents/skills/augustus/SKILL.md @@ -3,7 +3,7 @@ name: augustus description: "Use when placing typed probabilistic judgment (Jev-class System One / decision models) with mathematical, logical, or algorithmic mental models — in AI, software, business, knowledge work, or life, not only SWE; deciding where a fast cheap categorization/classification/scoring model belongs versus generation, exact policy/code, or proof; applying expected utility, selective classification/abstention, calibration, cost-sensitive thresholds, value of information, MCDA, signal detection, search/control substitutions, or Leveson-style org/safety; using NATM/snap-fit/Norman as design intuition; designing mixed architecture (decision model + LLM writing); auditing an existing system, PR, workflow, or non-software practice for judgment-shaped holes and code smells; debugging a question that hovers near 0.5, clusters mid-scale, or hides two judgments; placing agent self-supervision gates (pre-action, output judge, done-check, stuck-detector, context sieve); coupling a typed judge as an optimizer metric (Ax, DSPy); choosing among TypeSafe Jev, open heads (Laya, kev, openjev-lm, Nimble, encoder DeBERTa, LoRA distill, open multimodal RLCD / blackwood), announced open decision-model (Watch — still not landed), constrained-AR (TypeAR, pcdServer, decision-token LoRA), diffusion structured reads, GLiNER/GLiClass/GLiGuard encoder family (locate vs categorize vs safety-schema classify vs local multi-head), listwise rankers, or vision scorers; placing judgment beside TLA+/Alloy/Apalache/Dafny/DST (Antithesis, Resonate, PufferLib) without laundering a Noul as a proof; answering \"it's just classification\", \"is Jev probabilistic programming\" (marginals vs joint, not a PPL), \"low/medium/high entropy\" (allocator, not a meter), \"perception specialist then judgment vs shared multimodal System One\", \"wait for Archer vs open multimodal RLCD\", \"screenshot/DOM candidates → typed Choice\", \"eval path\", \"jevals\", \"Harbor taskset\", \"shared bake-off ECE/NLL/Brier\", \"LLM-as-judge is not the System One score\", \"pipeline / measure / hill-climb perception into a decision\", \"Ax vs DSPy\", \"held-out\", \"correctness is not confidence\", \"is this only for software?\", Alloy vs Apalache, GLiNER vs Jev, \"is GLiGuard Jev?\", LLM-as-judge, paraphrase brittleness, allowlist then judge (allowlist *proves*; fail-open cannot block), \"missing other → confident wrong Choice\", \"lint the Jev request\", \"training confronts Choice other / none-of-the-above\", \"S1 reflex keeps control / optional S2 one-use advice\", \"soft AGENTS.md rules vs linter\" (Abide / jev-pref), \"edit-phase vs turn-phase observation window\", banded confidence fail-open preference lint, \"extractive selection / pointer-not-generator\", \"encoder GLiNER compaction vs Jev Score compaction (same job; pointer not summarizer)\", \"fail-closed keep_full under mutation envelope\", \"CI flaky-vs-real merge gate\", \"fail-open VOI wake/resume (Horvitz)\", \"claim/evidence Stop integrity\", \"S1 extract + escalate-S2 indexer\", \"Harbor on/off routing\", \"policy-as-judgment PR marshal\", \"shadow-mode compaction rollout\", \"Jev Ultrafast vs GLiNER Ultrafast (observe-score-act backend-agnostic)\", \"hybrid local decide + remote fill\", \"DONE ≠ verified success\", \"observed a11y/DOM candidates vs screenshot multimodal\", \"evidence-preserving stdout prune (not summarize)\", \"hard token/format envelope then soft Noul\", \"fail-safe keep original on prune failure\", \"stdout prune vs session compaction\", \"specialist S1 computer-use (Cua-S1 form-v0; not TypeSafe Jev)\", \"plan ≠ execute / dry-run default\", \"observed-element option head (fill/check/click/skip)\", \"local /v1/systemone drop-in (stub until hf scorer)\", \"dataframe-native semantic columns\", \"route≠memory\", \"advisory sidecar receipts\", \"structure induction over bags\", \"AST ∩ semantic lint\", \"extractable-from-state / retrieve first\", \"decision-model vs constrained-LLM bake-off\", \"dual-process S1 decide / S2 generate\", \"combinatorial grid ≠ extractive\", TOCTOU-of-Noul, vacuous specs, open weights vs constrained decoding vs encoder vs LoRA vs kev, whether a decision needs a model at all (meta-VOI), env-break vs policy-break, sqlite-jev / in-engine vs CLI store index, hard safety envelope (Jev proposes, code clamps), host-adapter routing (not MCP), distill-to-device memory gate, \"uncalibrated local likelihoods vs Noul / CUDA replica\", \"decision-native RAG retrieve wide then decide then evidence set\", \"classify-first MCP / read selectively\", \"living applied-mappings atlas / class patterns not a 342 hit list\", \"draft-gate silence as safer / heartbeat\", \"robotics text-state vs pixels\", \"verbatim session ledger / scored recall\", \"judgment as language primitive / English-as-config\", \"pre-registered AMBIGUOUS eval / cascade sign-flip\", \"healthcare Harbor-shaped S1+S2\", \"pre-exec tool gate allow/block/review\", \"productized public primitive / judgment wall\", \"meaning-search without embeddings\", \"attention≠correctness PR review\", \"skills→oxlint / AST prove ∩ remainder\", \"session-sticky first-prompt routing\", \"measured RAG rerank vs generative rerank\", \"Stagehand extract pick-and-copy / judge\", \"harness observe-score-act productization\", or \"formally verify with Jev\", \"capability kernel / secrets never in the agent\", \"Jev is SENSOR not policy\", \"type-safe ≠ correct\", \"typed control plane around DSPy\", \"native-probability calibration / Brier/ECE arena\", \"fan-out as measurement economics\", \"engine owns truth / Jev owns judgment\", \"human-confirmed kill gate\", \"train specialist when downstream reads p vs few-shot hosted when only argmax\", \"decide→policy→LLM leftover cascade\", \"Noul 0.5 cannot-tell never rounded\", \"calibration ≠ sortable / ORDER BY over Jev probs\", \"pairwise inversion / Score ordinality / two-decimal ties\", \"wire-compat self-hosted /v1/systemone GLiFormer\", \"class-backend economics\", \"loopback gateway hosted + local OpenJev\", \"do not distill Jev as teacher of record\", \"active-learning triage / training-data VOI\", \"index-once ask-many / citable evidence packets\", \"meaning-grep AND/OR/NOT line Nouls\", \"closed-vote-only computer-use / no planner LLM\", \"Jev vs local MLX PCD Harbor\", \"PCD O(1) speed ≠ calibrated Noul\", \"host-owned handlers × System One\", \"OMP/pi fail-open acceptance gate\", \"permission vs probability / operator owns the safety bar\", \"judgment ≠ permission / Jev never grants access\", \"eval integrity / instrument not score / dinostomp jev-as-if\", \"constrained optimizer + S1 features / never sole hot-path gate\", \"privilege ≠ verdict / effect contracts not tokens\", \"attention filter / VOI for human review / never blocks / never green unless sure\", \"measurement owns endorsement / evidence-gated question packs\", \"Jev supplies evidence / code owns authority\", \"ranking ≠ calibration / never hard-threshold raw p as frequency\", \"hot-click CU / indexed element table / S1 on click path\", \"Jev judges relevance / code decides structure / never rewrite\", \"local rules first then remainder / never auto-train on model's own hides\", \"combinators / System One as control plane / not chat turns\", \"receipts not leaderboard / type-safe ≠ correct jaggedness\", \"VOI over skill library / skillranker abstention\", \"OOD calibration / AUC ≠ ECE / sign of miscalibration by type\", \"Jev vs thinking-budget small models / frontier-100\", \"turnstile / replayable evidence≠authority\", \"MLX one-pass schema→JSON / Apple Silicon replica economics\", \"memory leases ended by new evidence\", \"never confidently wrong / TLA+ compose with judgment / escalate instead of hard-gate\", \"no seal no advance / coverage ledger / mint ≠ product brain\", \"skill-broker sibling turnstile/skillranker / judgment ≠ permission\", \"sureness / CERTAIN|CONFIDENT|LEANING|TORN|CLUELESS / max_prob is generous\", \"JevBench Harbor/jevals practice / calibration not in Main Score\", \"CI typed gate before expensive review / ci-gatekeeper\", \"Codex MCP host adapter / jev_select_capability\", \"judgment as attention redirect not merge blocker / jev-preflight\", \"compress-before-first-send / dizk jev-lens vs rashed attention filter\", \"tools≠use / SessionStart over hoping the model recalls\", \"observational memory / keep-kind verbatim / pi-om\", \"open-Jev class / openvons / JevPick menu decode\", \"physical-world System One / HA-Jev / not for locks\", \"judgment outside the store / jevql CLI\", \"landed-script trust / headless≠auto-approve\", \"digital-design combinators / extended Router Loop Retry Fallback Memory\", \"VOI cache admission / same-intent skip LLM\", \"BM25 vs Jev skill routing Harbor harness\", \"zeroshot vs BERT / contamination DiD / label-equivalence\", \"typed escalate continue abort baton / inverted loop\", \"worth-your-attention VOI / ThinkyMiner Winnow vs kevinpita winnow\", \"Jev WHETHER Python HOW LLM WHAT\", \"conflict vs ignorance / named Choice escape\", \"Playwright executes Jev chooses / sample-from-distribution\", \"OpenJev /v1/decide not TypeSafe drop-in\", \"SemIf wire-compat runoff\", \"decision-as-memory flywheel\", \"record/replay CI / jevassert\", \"failure-finding arena / jevarena ≠ jev-arena\", \"BBQ stereotype/uncertainty/cost\", \"decider≠executor / jeffrey\", \"sentence-as-rule lint / jevlint ≠ JevLint\", \"VOI hunk prune / prune-review\", \"whole-repo intent VERIFIED/VIOLATION/UNKNOWN\", \"GLiNER2 System One spec ≠ replica\", \"Rust/WebGPU grande / Clojure Laya byte parity / CPU SemIf\", \"ONNX ModernBERT local-jev measured not equivalent\", \"persist constraints across compaction / pi-heed\", \"calibration+cost as first-class gates\", \"Harbor-shaped Jev vs schema-guided LLM-as-judge / jev-judge-bench ≠ jevarena ≠ jevbench\", \"hand no-text steps to Jev / jev-use / Vercel drops confidence / margin fallback\", \"Pi System-One control plane / pi-jev-control\", \"generation as tree of Choices / never free-generates / jev-gpt\", \"OpenRouter recipe atlas / samples not benches / jev-cookbook\", \"personal history feed / no social graph / jevfeed\", \"competing NAR claims / dual-channel ECE / claim-verification / openJev-verdict ≠ OpenJev\", \"empty compaction-proxy skip / IPECTER\", \"throughput ≠ latency / like-for-like ECE\", \"1-token logprob endpoint ≠ Noul / coverage ≠ correctness / chakuho\", \"open replica engine / jevinf / argmax-parity ≠ ECE\", \"unofficial Elixir SDK ≠ OTP peer / dannote/jev\", \"jevex rename + n=16 SWE VOI / files-to-read\", \"commit pre-review attention≠verdict / middle band never rounded / commitjev\", \"Hermes plugin is Agnes not TypeSafe\", \"pi-jev-compact ≠ pi-jev-compaction / verbatim summarizer replacement\", \"empty Codex-proxy skip / IPECTER runway\", \"decision-native inbox / mailordinal / humans own ambiguity\", \"unofficial jev-cli not ready / ≠ jevql\", \"laya-multilingual / English checkpoint confident-wrong OOD / ships uncalibrated\", \"schema-conditioned DeBERTa scorer / peaked ranking ≠ calibration\", \"HF 401 access / GitHub 404 Hub-only\", \"productized System One HTTP / classifier.dev / label+confidence public contract\", \"escalate-under-threshold / smart tier 0.7 / multi-label ignores tier\", \"silent-fallback FALLBACK marker / granite 0.546 vs advertised 0.800\", \"vs_jev tracked JSON not transcription / read eval/README before quoting\", \"choxos/jev-reviewer ≠ egma-ai / systematic-review pointer-not-generator\", \"two-pass Choice+Noul / relative which-line + absolute does-this-line\", \"not-found is an answer / no paraphrase invent\", \"human check as productized judgment / checked never overwritten\", \"githubnext/localjev ≠ kunchenguid/local-jev / prompted JSON ≠ structured logit read\", \"wire-compat ≠ logit-equiv / self-reported probs / entropy confidence\", \"institutional open-replica / GitHub Next /v1/systemone\", \"Harbor-shaped bake-off AG News BoolQ SST-5 / 1200-request caveats\", \"LM Studio runner gap / structured-read primitives for OpenJev parity\", \"NandhaKishorM/laya packaging ≠ Hub-only / Router script-before-p\", \"post-T ECE ≠ raw ECE / Banking77 token-budget / 0.85 still soft / not TypeSafe drop-in / external census ≠ scored bake-off / GLiNER2+routers class-boundary / incomplete vs watch / Harbor honesty watch / JevBench v1.2 geometric-mean I/C/S/K / cal now ON rank / weight sensitivity / option-order 72→21 / instruction models class-boundary / ×2 latency assumption / est. costs / Laya absent gap / Qwen3.8 27B ≠ Archer\", \"hourly already-folded watch / apply-the-five / skip thin noise\", \"hard-gate Noul as PR gate is soundness theater / totally-tim/jev-gate ≠ jev-gateway\", \"S1 never stalls waiting / S2 one-use advisory\", \"purple telemetry = consumed not arrived\", \"Local controller ≠ githubnext/localjev\", \"seed = geometry not async replay\", \"20% starting gate still soft / schema-safe ≠ correct\", \"no pixels to either provider / confidence ≠ selected probability\", \"experimental viz not a flight controller / S2 never grants\", \"OCR+AX observe-score-act / typesafe-computer-use\", \"never send screenshot to frontier for the decision\", \"overlapping CU options = false low confidence\", \"split kind/item/site / offscreen\", \"writer/decider split + post-type Noul still soft\", \"155× one-screenshot Harbor-shaped ≠ taskset\", \"AX never sole / Spotify 0\", \"decision ≠ answer-reader capture\", \"typesafe-computer-use ≠ jev-ultrafast ≠ cua-s1 ≠ camoufox\", \"ASR observe-score-act / jev-voice-browser\", \"partial-speech VOI / complete Noul / free-text waits\", \"spoken confirm ≠ hard auth\", \"numbered overlay disambiguate without another model\", \"moritzkremb/jev-voice-browser ≠ jev-voice-control ≠ nikolas-j\", \"wrap-as-execution / AgentGhost ALLOW ASK DENY\", \"rules first then Jev remainder / ASK throws / fail-closed\", \"reddpy/AgentGhost ≠ jwen5419807/agentghost ≠ vventirozos\", \"JP genre atlas / studio_yebisu / stars ephemeral ≠ eval\", \"Jev Clearly Explained / akshay_pachaar / LLM hammer\", \"schema-safe ≠ correct / 200× 400× TypeSafe ceiling\", \"questions-as-code / shadow first / not a TypeSafe how-to\", \"proposition ≠ embedding / contrast-set refund\", \"boolean composition of soft Nouls / AND OR NOT after threshold\", \"uehaj/jev-semgrep ≠ semgrep.dev\", \"meaning-grep dedicated fold / not a gate\", \"decision-validated UI / Jev never authors text / gram-render\", \"decision-as-assert / jevtest ambiguous band\", \"typed decisions drive UI / jev2ui\", \"hybrid S1 closed verb menu / anima3 / jeff confidently flat\", \"pointer-not-generator search / JevFind\", \"jev-frontier-bench ≠ frontier-100 / ChaosNLI JS\", \"product bakeoff ≠ architecture duel / jev-gliclass-bench\", \"four engines same questions / majority floor / calibration ≠ discrimination\", \"authorship named escape / not courtroom evidence\", \"ha-switchboard HA remains execution / ≠ HA-Jev\", \"n8n classify/route/score / Low Confidence abstention\", \"fast-jev-compaction-pi ≠ pi-jev-compact ≠ pi-jev-compaction\", \"jevloop full-distribution optimizer / no LLM in the loop\", \"laya-vision SmolVLM / score untrained / ≠ blackwood ≠ Archer\", \"Cerebellum-2B /v1/decide ≠ TypeSafe / wire-compat vs agent-routing\", \"laya-grounded not drop-in / phishing regress / Platt not temperature\", \"GestaltLabs/Jeff-1 ≠ logan-markewich/jeff / acc vs ECE n=9730\", \"stanley-code empty findings ≠ approval / human promote\", \"findme ≠ JevFind / NL memory beam-search FS\", \"jevsubrouter price workers not conversation / counts ≠ dollars\", \"feelings .feels() default 0.5 is Noul-0.5-never-rounded / ≠ hunch ≠ Probably\", \"apa-agent-harness ≠ AntonioCoppe/jev-harness / unpublished npm\", \"grok-bot-jev skill cannot force a bot that ignores it / A/B proxies not tokens\", \"Essentiel-Jev never authority / human every action\", \"enzo-mcp independently falsifiable claims / ≠ jev-sift\", \"pigeonhole OTHER skip / decision-as-filing\", \"jev-reliability Nothing about accuracy\", \"clduab11/jev-test ≠ realZachi/jevtest / Nothing runs yet\", \"jev-rag-benchmark Jev wins is not an assumption\", \"dairui1/jev-lab ≠ BrendanH18/jev-lab\", \"jevmail gmail.readonly / mailjay archive/trash\", \"ZHUBoer/ego-jev reserved __none__\", \"runWorkflow completed ≠ success\", \"jsort scores are relative\", \"Noul not Choice for scale\", \"groundedness-judge-bench native vs schema-guided\", \"implicit_true included in yes\", \"jev_playground 0 promotions\", \"routing-backtest 0.0447%\", \"yuyang2230/jev-agent-skill jev-1.13-free\", \"jev-techstack-classifier stack_config.json\", \"s1_ruby collapse late\", \"undecided? abstain\", \"2389-research/judgement license null\", \"confidence ≠ winner p\", \"typesafeai-sdk-community not a new species\", \"tpellet/hunch exit 3\", \"never-execute list\", \"jev-file-search scores not calibrated accuracy\", \"jev-linkmap Jev never sees S2 prose\", \"muhammedilyasy/jev-mail metadata only\", \"tidy none-of-folders stay\", \"tab-bouncer pinned/audio/current never closed\", \"lkclean Show fail-open\", \"jev-yt-time-saver Show anyway\", \"ORIGIN pause-if-no-Jev\", \"validResponse sums-to-1\", \"jev-crawlers risk bands never raw boolean\", \"jevbrain AUTO_ACT is not a Noul\", \"judgekit YAML classify/score/route/verify\", \"typed-judge-kit verdict-in-code\", \"alsoleg89/decide packing VOI\", \"0.8 ≠ 80% accuracy\", \"Jev-Calibration Platt ECE 0.117→0.052\", \"jev-calibration-arena never acts\", \"ctmx/openrouter-jev-mcp Decision-as-Plugin\", \"FrancoisChastel/jev-code ≠ npm jev-code\", \"claudecode-jev-marketplace fail-open not hot path\", \"pedroknigge/mcp_jev packs not ask_jev\", \"cyrusasco/typesafe-mcp noul deadband 0.35–0.65\", \"codaaiteam/jev-skill jevtypesafeai.com ≠ TypeSafe\", \"hermes-switchyard ≠ hermes-jev-router ≠ hermes-plugin-jev\", \"nanoprune 2.8MB ECE 2.58%\", \"smartdio/jev-browser-agent ≠ ZHUBoer/ego-jev\", \"Dakai/omp-jev-web DONE ≠ proof\", \"hari007sh/jev ≠ dannote/jev\", \"0thernet/system-one-skills deterministic verify\", \"typed-gate band [0.40,0.60] is refusal\", \"pi-jev-gate fail-closed; choice is the verdict\", \"Foq ~25ms/2.2GB local\", \"rev prefill-only + HF jev-0.5b\", \"robfrase/jev planning memo\", \"typesafe_agent_gates 27/27 / 31/31\", \"EpicEric/safe-sh static remainder\", \"pastepilot Confirm before act\", \"Jev-Reranker live Jev not yet measured\", \"sessionwise opt-in relevance\", \"jev-search pointer sieve\", \"400ms Salesforce WebMCP\", \"typesafe-scheduler-diagnostics advisory\", \"droidjev screenshot-free\", \"Tewoto1 jevcu planner still writes\", \"ha-conversation-jev Jev→Grok\", \"dsh-jev can only gate\", \"jev-classification-benchmark specified not run\", \"jev-luna-pagerduty p≥0.50\", \"meldltd/meldecision laya-go ONNX\", \"laya-doom never pixels\", \"logixism/laya-api empty README\", \"akpsahan/laya ≠ Archer\", \"choxos/jevchess engine owns truth\", \"jev-drive sim not AV\", \"story-arc Jev never authors\", \"jev-hs-assistant HS6\", \"golergka/jev-plays-starcraft-2 UI-verified ≠ API Victory\", \"awesome-jev-use-cases catalog\", \"Nibir1/typesafe-go ≠ official\", \"fingerprint after redact\", \"recall vs decide\", \"publish fingerprints+answers\", \"CI replay as Harbor cousin\", \"Cache hit ≠ correctness\", \"hyperspaceai/jevcache ≠ kushals256/jevcache\", \"human labels only\", \"score never auto-accepts\", \"production capture flywheel\", \"sutro-sh/jev-align ≠ caiovicentino/jev-align\", \"guidance ≠ hook\", \"catalysts ≠ summaries\", \"compile-time System One\", \"unofficial ≠ TypeSafe\", \"format_version modernbert-jev/1\", \"Argos1111/jev_local ≠ us/jev-local ≠ kunchenguid/local-jev\", \"LFM default ≠ ModernBERT backend\", \"Nemotron ≠ TypeSafe Jev\", \"not a calibrated replacement\", \"djev-dev complements djev-spark\", \"images as Choice options\", \"Laya essay numbers *theirs*\", \"Router/OOD confidence\", \"hosted bootstrap ≠ silent TypeSafe\", \"difficulty + policy thresholds + JSONL trace\", \"jev-codex-pilot model + reasoning depth\", \"keep/shadow/hybrid/reject\", \"quarry evidence projection\", \"Frank-ZY-Dou/awesome-jev robotics/3D/control\", \"one-dollar-tahoe TypeSafe Jev defense eval\", \"jevguard calibrator/cache/escape\", \"jev-ci-selector CI shadow mode\", \"llama-jev llama.cpp replica\", \"petercr/jev-orchestrator ≠ FleeexCorp/jev-orchestrator\", \"seb4ez/jevguard ≠ AseemPrasad/JevGuard ≠ pablozr/JevGuard\", \"webNeat/llama-jev ≠ WiktorB2004/llama-index-jev\", \"OpenCode jev-pruner context sieve\", \"observe→score-candidates→prune\", \"jev-zen / jev-1.13-free\", \"zen-chat ≠ Noul\", \"fail-open original\", \"keepScore >0.1 floor\", \"host port of tamaratran/jev-pruner\", \"indiejoseph/opencode-jev-pruner ≠ nrdz-labs/fast-jev-opencode\", \"jev-webagent-bench empty stub\", \"Kiln-AI/jev_jsonschema noul_threshold 0.5\", \"NSStudent/JevSwiftSDK unofficial\", \"GLiNER2 native Apple path\", \"unofficial Swift/Core ML GLiNER 2.5-small\", \"entity spans + confidence\", \"not Choice/Score/Noul\", \"not TypeSafe\", \"label descriptions as schema\", \"on-device ANE economics\", \"honesty locks\", \"shershah1024/gliner-native-runtime ≠ Fastino\", \"≠ gliner25-compaction ≠ gliner2-ultrafast ≠ Eran-BA/Jev_from_GLiNER2 ≠ NSStudent/JevSwiftSDK ≠ jevmlx\", \"default threshold 0.1 still soft\", \"Decision Graph Protocol frame→assess→commit\", \"app retains permissions/effects\", \"Jev-first assessor-neutral\", \"guarded commit / receipt/next frame\", \"assessment batching\", \"hard-gating DGP as safety theater\", \"numerous-com/dgp ≠ TypeSafe official\", \"jegrep calibrated path+range Nouls\", \"no embeddings/index/daemon\", \"~$0.01–0.03 typical\", \"agent --json\", \"can1357/jegrep ≠ Bentlybro/jevgrep ≠ uehaj/jev-semgrep\", \"Archer-arch fidelity\", \"kev family OOD 0.76–0.77 vs Jev 0.86\", \"block-causal isolation\", \"pointer/readout CE-trained\", \"/v1/systemone drop-in\", \"replica honesty\", \"cost-sensitive decision theory × System One probabilities → control flow\", \"thresholds derived from costs not hard-coded\", \"YES / NO / UNSURE from cost_false_yes / cost_false_no / cost_human\", \"auto-batching same-object questions\", \"Kungie/gut ≠ tpellet/hunch ≠ carldaws/hunch\", \"judgment vs generation\", \"deterministic execution after probabilistic judgment\", \"exactly one app-owned callback\", \"explicit uncertain branch\", \"Illusion47586/judge ≠ lexingtonhibiki/judgekit ≠ Ascurse/typed-judge-kit\", \"variable-N option scoring as the trainable object\", \"dynamic candidate bags not fixed label sets\", \"zwliJay/jev-forge ≠ NanoJev\", \"open replica economics / latency vs closed Jev\", \"NAR local drop-in\", \"wfzyx/von late-catch HIGH\", \"competing NAR claims / replica honesty\", \"typed judgments vs chat judges on guardrailing\", \"ishaannk/llm-vs-jev cross-note only\", \"deeper integrity fold is rh-guard\", \"nothing wins outright\", \"can be argued out of guarding\", \"Jev IS the if-statement\", \"judgments/probabilities drive branches\", \"text model only writes prose\", \"interpreter owns variables/loops/budgets/replay\", \"otherwise maybe / confidence gate\", \"chaos samples after the gate\", \"southpolesteve/probably ≠ carldaws/hunch ≠ feelings ≠ Kungie/gut ≠ Illusion47586/judge ≠ tidymodels/probably\", \"133★ / forks 10 live\", \"build calibrated classifiers from human feedback\", \"retrieve by relevance not resemblance\", \"one calibrated yes/no per memory in one request\", \"pointer mode 17/18 19/20 *theirs*\", \"embedding resemblance misses the allergy\", \"samdotmak/jev-recall ≠ jev-search ≠ jev-sift ≠ carryforward ≠ chopratejas/invalidate\", \"memory leases ended by new evidence\", \"six Nouls then fixed rules in code\", \"0 of 157 false invalidations\", \"questions/plans/directives are not evidence\", \"unsure → review queue\", \"host keeps the store\", \"name↔body / comment truth / test-claims\", \"mizchi/jev-lint is mizchi/jevlint rename\", \"no shipped rule has severity error\", \"~1 in 5 findings wrong *theirs*\", \"mizchi/jev-lint ≠ huntedman/JevLint ≠ MichitoSugawara/jev-lint\", \"JSON Schema → typed JSON via Jev\", \"noul_threshold 0.5 decoder not a proof\", \"IncompatibleSchemaError lists every bad property\", \"on-device Laya CoreML ANE\", \"~5 ms P50 short decisions\", \"189/189 FP16 checkpoint parity\", \"10× not achieved\", \"mizorewww/laya-coreml ≠ gliner-native-runtime ≠ jevmlx ≠ NandhaKishorM/laya\", \"softmax over allowed tokens ≠ Noul\", \"question-first cache\", \"Micha0827/snapjudge ≠ githubnext/localjev ≠ jevmlx ≠ cendress/SnapJudge\", \"Jev-first Pi agent loop\", \"slow-LLM fallback\", \"explicit action menu / CandidateSource unimplemented\", \"62 tests wiring not quality\", \"direwolfiy/JevPi ≠ standardagents/jevpilot ≠ pi-jev-control\", \"resume-screening bias audit methodology\", \"name×resume factorial independent Nouls\", \"callback determined by resume quality\", \"mean-probability name gaps operationally negligible\", \"natemoo-re/bias-bench ≠ BBQ\", \"Plan/PRD panel → code-owned pass|review|block\", \"cheerleading out of scope\", \"austindixson/planalyzer ≠ single-goodness Noul\", \"cost-aware multi-model routing/escalation\", \"decide vs do\", \"successful-task cost\", \"cannacre8ive/switchboard-ai ≠ ha-switchboard ≠ hermes-switchyard\", \"frozen-protocol zero-shot bench\", \"TypeSafe Jev vs PrismNLI vs Laya\", \"contamination caveat\", \"elcronos/jev-vs-open-decision-models ≠ JevBench ≠ DMB\", \"context-window admission control\", \"VOI gate which tokens are worth the expensive model\", \"fail polarity per lens\", \"on small inputs lenses lose money\", \"cvsgireesh/jevusher ≠ jev-sift ≠ winnow\", \"typed decision control plane\", \"receipt ≠ authorization\", \"historical-v0 zero retained cases\", \"MokiMeow/jev-fabric ≠ jev-forge ≠ dgp\", \"live 15-dim typed rubric re-score per pause\", \"scoring economics exemplar\", \"OpenJev/Codiv ≠ TypeSafe hosted\", \"jose-troche/live-rubric ~$0.000004 desc / ~$0.000006 README\", \"adversarial pre-registered Jev eval\", \"28 predictions before data\", \"123,805 requests\", \"confidence does not track ignorance\", \"polite injection 65% / crude 0%\", \"willkelly/jev-evaluation ≠ jevals ≠ jev-baselines-eval\", \"provider-neutral Elixir/BEAM Noul/Choice/Score SDK\", \"class infrastructure\", \"nshkrdotcom/system_one_sdk ≠ typesafe_sdk ≠ dannote/jev\", \"question-linting of Jev questions themselves\", \"nine jaggedness rules, no API key, no labelled data\", \"static lint ≠ measured separation\", \"yodablocks/jevq ≠ tenbin ≠ JevLint ≠ commitjev\", \"open-weights Laya as class exemplar (binding)\", \"Nx/Bumblebee runtime\", \"host chooses backend\", \"ChristianAlexander/laya_ex ≠ system_one_sdk ≠ dannote/jev ≠ NandhaKishorM/laya\", \"on-chain/edge Laya deploy\", \"parity_verified stays false\", \"model output never grants Tx\", \"humandebri/IC-Laya ≠ laya_ex\", \"auditable weekend replica\", \"Jev outputs never used for training\", \"soft human-vote distributions\", \"unpaired 0.577 vs 0.727\", \"agilabs-ai/jev48 ≠ JevBench ≠ Mapika/decider\", \"adversarial dual-judge / framing attack surface\", \"comparative framing is the usable judgment\", \"prior injection crowds out evidence\", \"copyleftdev/ember ≠ ember.js\", \"Laya specialist fine-tune pipeline\", \"training still GPU-pending\", \"PIXELZX0/XERON ≠ convaiinnovations/laya\", \"Hub Laya replica drop\", \"daliborsb/laya ≠ convaiinnovations/laya ≠ NandhaKishorM/laya\", \"System One student distillation corpus\", \"gold is programmatic\", \"teacher is closed-API clone\", \"do not distill Jev as teacher of record\", \"MagaBitmex/jev-4b-distill-data ≠ missing student checkpoint\", \"non-LLM VIN System One\", \"planning depth not chat\", \"lewislululu/jevon ≠ douglance/jevon\", \"source-bound evidence checks\", \"local quote mismatch needs no API\", \"exit 0 ≠ claim truth\", \"WaynezProg/jev-kit ≠ jonathanavis96/jev-kit (Airlock) ≠ jev-use ≠ jev-mcp\", \"independent System One evidence catalog\", \"scores not one leaderboard\", \"no external record currently reproduced\", \"TokenTrim no-Jev matched hybrid 62.4%\", \"reachjalil/system-one-bench ≠ mallahyari/system-one-benchmark\", \"21 tasks · 134 items · 208 questions\", \"scenes from public GitHub contracts, not production logs\", \"SivletLabs/jev-eval ≠ willkelly/jev-evaluation ≠ 4esv/jev-eval ≠ xxkuboxx/jev-eval ≠ onlyoneaman/jev-eval ≠ dayhaysoos/jevals\", \"option isolation (sibling-blind)\", \"permutation-equivariant\", \"Hub OWNER not published\", \"nafisazizir/hev ≠ jaredpalmer/kev\", \"frozen local LLM logits, no trained decision head\", \"residual-head 9,222-param decreased 73/96→67/96\", \"confidence = 1−normalized entropy, not P(correct)\", \"yuki-oshio/mini-jev ≠ r-ms/mini-jev\", \"Jev classifier as autoregressive next-token predictor\", \"ChatJev-style soundness theater\", \"erik-dunteman/ChatJev ≠ dannote/jev ≠ jev-gpt\", \"calibrated decision head × AlphaProof value head\", \"implementation-layer isomorphism, semantic difference\", \"timeout = censoring\", \"do not launder Noul as proof\", \"parallel rank-prediction vs serial selection\", \"independent questions can conflict\", \"zzzzzec/jevsort ≠ keltokhy/jsort\", \"curated open System One ecosystem catalog\", \"rupeshpoojary9/awesome-open-system-one ≠ AnotiaWang/awesome-jev\", \"arXiv paper radar with Jev relevance scoring\", \"ranking ≠ calibration / 0.5 still soft\", \"fail-open failed evals not marked seen\", \"train calibrated ~27M from scratch\", \"typed Q→prob dist / one forward pass / no LLM decode\", \"hyusi2003/MiniSystemOne ≠ Colvin0315/MiniSystemOne\", \"description-only stub / size 5\", \"ESCI hard probe fails four of six\", \"jev_bool ECE 0.242 inversion 0.255\", \"do not re-fold §60 six-gates as new\", \"jobbyjev one-request-per-company from batch-size result\", \"find/design/evaluate TypeSafe Jev decision loops\", \"karanb192/jev-architect ≠ samtay32/jev-system-architect\", \"Jairik/jev-distiller size 1\", \"distill-Jev UI stub / do not distill Jev as teacher of record\", \"post-launch scored use-case map / Jev self-scores then human curation\", \"licensedsaucer9-web/jev-opportunities\", \"Jev-inize a use case into classifier/router\", \"gavinHuang/jevinize → simple-jev not TypeSafe\", \"featherless-ai/simple-jev\", \"compare saved decisions / same label can still change the branch\", \"VihaanAgarwal/jev-diff ≠ Saik0s/diffusiongemma-jev-macos\", \"not tested with a live Jev API key\", \"constrained logprob + temp/Platt ≠ Noul\", \"OpenJevPro pastes openjev-sglang JevBench as own\", \"zhangcy122/OpenJevPro ≠ IamBusy/OpenJev ≠ ekzhang/openjev-sglang\", \"PolyForm Noncommercial\", \"SmolLM-135M / sub-70ms / 0 output tokens\", \"demo P(True) 0.5052 / Choice conf 0.2872 / Score conf 0.0055\", \"README claims MIT / GitHub license null / no LICENSE file\", \"patelvishwa112/jev-system-one-rlcd ≠ arnabgho/rlcd-lite ≠ blackwood-rlcd\", \"source-backed Awesome Jev radar / 306+ commit-pinned\", \"logicrw/awesome-jev-projects ≠ AnotiaWang/awesome-jev ≠ yibie/awesome-jev ≠ cobanov/awesome-jev ≠ rupeshpoojary9/awesome-open-system-one\", \"auto GitHub sync / Issue-only submissions\", \"hashed n-gram encoder / rival-aware attention\", \"olanotolu/jevbetter vs jevlike starter\", \"synthetic hard menus top-1 0.916 vs 0.873 / ECE 0.0182 vs 0.0367 / 40 vs 4608 menus/sec\", \"shuffled-context control 0.335\", \"structured probability readouts\", \"distribution > argmax\", \"Noul 0.5 midpoint\", \"score is expectation not integer\", \"bare HTTP not SDK\", \"Arohtea/jev-readout\", \"Jev-style Choice/Score/Noul from ordinary models\", \"optional DSH plugin\", \"schema-valid ≠ calibrated\", \"gulagala001/jevify ≠ Mintzs/jevify\", \"Laya RLCD benchmark\", \"40.3% below constant-answer\", \"open-weight measurement\", \"mourad-ghafiri/laya-rlcd-benchmark ≠ yibie/laya-jev-lab\", \"cheap fail-open semantic edge\", \"second signal not sole\", \"FastLoopError catch\", \"SupremeDreamZ/jev-fastloop ≠ jev-ultrafast\", \"asking more questions in one call\", \"0.980 at every N\", \"nearly not fully deterministic\", \"TheWebDevel/jev-fanout\", \"Qwen3-VL perception + Jev decisions train RL\", \"0 model calls at deployment\", \"VLM alone 1.7 vs +Jev 4.4\", \"harneet2512/reflexrl ≠ khordoo/jev-reflex-autonomy-lab\", \"independent Jev API vs Laya\", \"cascade 0.60 matches 78% at 1.8×\", \"noul facts not judgements\", \"yibie/laya-jev-lab ≠ dairui1/jev-lab ≠ BrendanH18/jev-lab\", \"GLiNER vs GLiFormer vs Laya vs Jev\", \"extractors ≠ decision engines\", \"Laya dict-instructions collapse 58.3%\", \"umstek/zero-shot-ie-bench\", \"decisions-per-minute & cost\", \"204 moves vs 73\", \"throughput not intelligence\", \"angelgalvisc/snake-arena-jev-vs-llms ≠ vtrivedy/jev-plays-games\", \"behavioral contracts\", \"pin expectations eval upgrades\", \"raw 0.94 is not a release\", \"sathariels/jevcheck ≠ dayhaysoos/jevals ≠ SivletLabs/jev-eval\", \"evidence-linked dependency upgrade\", \"Jev never generates filenames\", \"no_direct_evidence ≠ safe to merge\", \"GaneshVG18/upgrade-radar ≠ LYchoon/paper-radar-jev\", \"discography theme/mood/complexity\", \"five atomic questions one call\", \"lirantal/discoprint\", \"Turn any open LLM into System-One Jev\", \"uspraveen/Jevify ≠ Mintzs/jevify ≠ gulagala001/jevify\", \"Jevify-any-LLM architecture probe\", \"description-only stub / size 0\", \"Train encoder-only calibrated decision models from a task sentence\", \"Exu is a toolkit, not a method\", \"strictly proper scoring rule\", \"Pre-alpha\", \"Ruivalim/exu-base\", \"scratch-trained calibrated decision model\", \"typed Q → probability dists\", \"Colvin0315/MiniSystemOne ≠ hyusi2003/MiniSystemOne\", \"no published weights download URL\", \"90.5 seconds / 29.2% pipeline evidence\", \"p_i/p_j independent of other candidates\", \"Recipe for calibrated decision models — small model out\", \"init → synth → train → eval → serve\", \"91.1 % / ECE 0.022 *theirs*\", \"Jev zero-shot 75.1\", \"scienthoon/luce\", \"Put Jev's three headline claims on trial\", \"0.5B local GPU\", \"46x speedup / accuracy identical\", \"ECE 0.624 sentiment catastrophe\", \"bigger model worse calibration\", \"RichardoMrMu/jev-mini ≠ yuki-oshio/mini-jev ≠ r-ms/mini-jev\", \"System-1 decision engine for local LLMs\", \"structured choices only\", \"JSON parse of generated text ≠ Noul\", \"TypefAI JEV / Journal Entry Voucher\", \"tapsin/jev-local ≠ us/jev-local ≠ Argos1111/jev_local\", \"Jev 1.13 reward-model eval across 8 benchmark tracks\", \"40,940 examples / 0 API errors\", \"RewardBench v1 92.58%\", \"Precise IF 50.63%\", \"goya4140/jev-reward-model-evaluation\", \"Scaffolding in progress\", \"Jev vs LLM support-ticket routing\", \"static + live decision bench\", \"TypeSafe's own published benchmark\", \"illustrative simulations, not live API calls\", \"JevBench v1 — smart/cheap/fast/reliable\", \"I/C/S/K 25% geometric mean\", \"classifier.dev fast tier 84.8 is Jev behind its own API\", \"do not re-fold §78 v1.2 board as new\", \"Laya (421M) 70.1 now on board\", \"Zero-shot/few-shot LLM routing\", \"hard budget filter before Jev\", \"Jev never asked to perform budget arithmetic\", \"Jev judges the next state, XState enforces transitions\", \"simulation uses synthetic keyword fixtures\", \"catalog gravity\", \"v-modal/awesome-jev-tools\", \"★339 live REST\", \"curation is not endorsement\", \"crawler-maintained directory\", \"Daily GitHub + npm sweep, human-merged\", \"RadRebelSam/awesome-jev ≠ AnotiaWang ≠ yibie ≠ cobanov ≠ logicrw ≠ v-modal\", \"HF peft SPLADE/BGE reranker\", \"rdxtremity/jev-reranking ≠ carlaiau/jev-reranking\", \"query-side encoders, not a Jev replica\", \"ONNX System One Qwen3.5-4B scorer\", \"source:pngwn/system-one-qwen3.5-4b-scorer\", \"CC-BY-NC-4.0\", \"temperature 1.75\", \"transformers.js AutoModel cannot load this graph\", \"Consistency benchmark Space\", \"This Space contains no benchmark result yet\", \"12-case plumbing fixture\", \"Benchmark-driven Jev router and judge\", \"cheap alone is not success\", \"Jev does not write, sum prices, or claim accuracy %\", \"Sol 94.2 / Luna 83.9 / Jev path 89.7\", \"19.2% Sol / 62.3% cost save / 4.5pp miss of 2pp non-inferiority\", \"p50 latency worse than Sol due to routing overhead\", \"erendikmenn/jev-llm-router-benchmark ≠ jev-rag-benchmark ≠ ryantsai/jev-llm-router\", \"Express + node:sqlite\", \"mock and Jev decision engines\", \"previous_ticket_count >= 3 is code\", \"MIN_CONFIDENCE 0.6 still soft\", \"substring false positives\", \"aesaganda/jev-ticket-router ≠ SarathChandraBellam/jev-vs-llm-ticket-router\", \"Universal Figure & Diagram Router\", \"confidence ≥ 0.85 hard-gate is theater\", \"generative AI banned from scientific plots\", \"six visual branches\", \"hoangngochuong24947-gif/jev-figure-router\", \"human-labeled (state, question, label)\", \"166,054 rows / 22 configs\", \"soft_label for human uncertainty\", \"Praveenrajus/jev-bench ≠ fstandhartinger/jevbench\", \"ternary bonsai System One GGUF\", \"openjev's mechanism, Bonsai's weights\", \"Hub does not ship weights\", \"100/100 easy T/F is not Harbor\", \"label_mass ≠ correctness\", \"stock llama.cpp Q2_0 silently gibberish\", \"NicolaiMTLassen/open-bonzi-jev ≠ NicolaiLassen\", \"transformers.js DeBERTa ONNX\", \"source:com-kotobalabs/open-jev-deberta-v3-large\", \"temperature 1.05\", \"AutoModel from_pretrained works\", \"onnx-community/open-jev-deberta-v3-large-ONNX ≠ system-one-qwen3.5-4b-scorer-ONNX\", \"107★ densify\", \"GH 151M vs README 149.6M\", \"PR #1 now closed unmerged\", \"do not re-fold §71 claim-audit as a beat\", \"typed decisions, RLCD, confidence-gated routing\", \"structured ≠ correct\", \"mock not live API\", \"26 tests\", \"wjdjdakf17/jev-study ≠ baekenough/jev-study\", \"bonzi-27b-v2 / ternary-8b / 27b-v1 GGUF family densify\", \"Hub still does not ship weights\", \"WANLI-256 74.6% / 65.2% / 71.1% *theirs*\", \"Bonsai 1 27B Q1_0 runs on stock llama.cpp\", \"ternary still needs PrismML fork\", \"hf:heman10x/openJev-verdict-2.0 twin tokenizer-only\", \"OpenJev Vision image classification + uncertainty\", \"CLEVR-4 held-out joint 0%\", \"hfdataset:IamBusy/OpenJev-Vision-Research-v0.1 12,832\", \"294,912 derived targets not independent samples\", \"Laya multilingual ONNX WebGPU typed-decisions port\", \"63/63 selected answers / 5.1e-4 CPU / 1.2e-2 WebGPU\", \"UpHash-Network/mini-jev is yuki-oshio transfer\", \"jev-injection-bench 11,900 labelled prompts\", \"Jev best ranking / Haiku better ECE 0.021 vs 0.058\", \"0.5–0.9 band is where Jev's numbers do not mean what they say\", \"Prompt wording moves panic 28%\", \"manojlds/jev-dspy-bench ≠ dspachos/jev-dspy ≠ jmanhype/jev-dspy-lab\", \"Jev agreement is similarity, never ground truth\", \"no aggregate quality grade or merge gate\", \"AbstentionBench-on-Jev rank 1 of 20 vs 2025 field\", \"question-asymmetry\", \"forward-looking 0.465 never extreme\", \"openkev calibration layer not a runtime\", \"ECE vs coverage independent\", \"select_threshold returns inf\", \"escalation catches uncertainty not ignorance\", \"misakaikato/openkev ≠ jaredpalmer/kev\", \"pdf-race Docling→Jev vs Gemini\", \"parser owns the wall clock\", \"12/12 tie is a tie\", \"titles selected not generated\", \"flopcheck 16 calibrated tweet judgments\", \"mechanical tells in code\", \"ZeroX-01/jev-atlas ≠ Zaious/jev-capability-atlas ≠ gorock007/jev-atlas\", \"catalog not endorsement\", \"Laya calibration lab Gradio MCP\", \"T never changes argmax\", \"confidence ≠ top-label p\", \"easy probe set refused\", \"40–48 rows too small to ship T\", \"Gemma-4 26B-A4B jevify classification+calibration\", \"LoRA adapter twin not independent eval\", \"Gemma-4 E4B jevify\", \"E4B LoRA stub card\", \"kushalpatil/jevify-gemma4 ≠ Mintzs/jevify ≠ gulagala001/jevify ≠ uspraveen/Jevify\", \"GH kushalpatil07/jevify 404\", \"PAWS 0.580/ece 0.288 is the weak cell\", \"smaller E4B slightly better OOD ECE than 26B-A4B\", \"Hub jevify merged LoRA ships weights\", \"bonzi Bonsai-8B v1 GGUF densify\", \"Bonsai-1.7B v1\", \"Bonsai-4B v1\", \"WANLI-256 64.5% / 60.2% / 52.0% *theirs*\", \"rank #4 / #5 / #6 of 6\", \"JulesHuisman/jev-eval scaffolding / README SHA c356a584 (was empty e69de29b)\", \"JulesHuisman/jev-eval ≠ SivletLabs/jev-eval ≠ willkelly/jev-evaluation ≠ 4esv ≠ xxkuboxx ≠ onlyoneaman ≠ dayhaysoos/jevals\", \"7 bands 6/10 vs 40 bands 0/10\", \"source receipts + confidence slider re-policy without re-inference\", \"32/32 synthetic is smoke not production\", \"classify HF datasets across typed semantic dimensions\", \"roadus2 watch misspelling\", \"lock roadius2/ultra_laya\", \"ultra_laya REVIEW defects\", \"default branch claude/laya-jev-review-gg5ppo\", \"XNLI EN 88.3% ECE 0.032 → RU 77.3% ECE 0.096\", \"Δ −11.0 pp [−14.2,−7.8]\", \"ECE +0.063\", \"MASSIVE no detectable difference at n=600\", \"confidence is function of p_max (r=1.000)\", \"pointer-not-generator 400 human-authored responses\", \"proposed ≠ authorized\", \"FewRel 160: Jev 85.0% vs lexical 13.125%\", \"gated 100% (95/95) coverage 59.375%\", \"J++ composable semantic computation language\", \"judge-jev 0.5 still soft\", \"947 repos scored\", \"A 273 / B 302 / C 372\", \"LLM rubric ≠ benches\", \"No benchmark winner is claimed\", \"phishing: naive 62.6% vs regex 91.8%\", \"5-atomic + LR 95.0% *theirs*\", \"AITuber tension ±15\", \"README npm global\", \"repo is Rust\", \"git-confess code owns counting/blame/ratio\", \"httpx exhibit 11% (13/119) *theirs*\", \"90d trend +12.40% vs random +12.75% vs BH +41.71%\", \"5m win rate 25%\", \"Awesomejev 656 entries / 38,160 stars\", \"tracker likes 64 (+4) lastModified UNCHANGED\", \"Laya present\", \"Blackwood ABSENT\", \"Archer still promised_not_landed\", \"Blackwood tracker ABSENT; likes 2 gated manual\", \"r = c - p_a\", \"ECE 0.021; acc 0.807 vs warmup 0.746\", \"calibration beyond ~500 tokens unmeasured\", \"Independent primitive\", \"11.57s vs 54.10s · 4.67× · 120/128 *theirs*\", \"default path is pretrained Gemma probs not trained RLCD head\", \"GH Meanblock 404; lock leesk212/JEV-CPU\", \"softmax over letter slots ≠ Noul\", \"WANLI 0.741 vs openjev v2 0.77 *theirs*\", \"3-way NLI ≠ Noul\", \"priority 0.464 = majority floor\", \"banking77 contaminated\", \"raw margins not probabilities\", \"GH jev-haiku-benchmarking 404\", \"do not distill Jev as teacher of record (they distilled Haiku)\", \"“0.9 is not one number”\", \"ranking ≠ calibration\", \"banking77 0.8–0.9 stated 0.86 actual 0.73 over-confident *theirs*\", \"≠ Praveenrajus/jev-bench ≠ fstandhartinger/jevbench\", \"$0.0000153–$0.0000226 vs circulating $0.0004 (~20×)\", \"Score is 0..n-1 expectation not 0–1\", \"Noul has no confidence field\", \"TCP floor 198.8 ms\", \"type reliability is not a reason to choose Jev (json_schema 5/5)\", \"gateway tax not one number\", \"Function-only 5/8 vs hybrid 8/8\", \"4/8 without Jev\", \"8 designed cases not conversion lift\", \"≠ RadRebelSam/awesome-jev\", \"200-row pilot Jev 86.5% 173/200 vs Gemini Flash-Lite 86.0% 172/200 vs Pro 87.0% 174/200 *theirs*\", \"not a ranking\", \"NLI Tetris argmax P(entail)−P(contradict)\", \"情緒測謊器\", \"1q 396ms / 30q 567ms\", \"±0.03\", \"33q $0.000045 vs Gemini ~5× slower ~60× cost *theirs*\", \"≠ realZachi/jevtest\", \"8-example Jev vs GPT-5.6 Sol ~64× cost 5.4× latency *theirs*\", \"synthetic; no inference\", \"≠ JevBench v1.2 §78\", \"Judged 3317 / listed 2560\", \"Jev judges, code applies policy\", \"catalog ≠ endorsement\", \"APA “microsecond policy / zero hallucination” overclaim\", \"Client-side quiz; pointer from held docs; scanned-PDF warn\", \"CSP only api.typesafe.ai\", \"Jev judges / agent reasons / user decides\", \"selecting an option is not permission to implement\", \"degraded fallback\", \"pattern exact, judgement must clear floor\", \"no matching pattern → no model call\", \"not a correctness oracle\", \"$0.00022 vs chat $0.00306 *theirs*\", \"Spec vs artifact remainder\", \"treating 0.85 as 85% / minProbability hard-gate as Harbor\", \"VERIFY acquires discriminating evidence, never same-pool confidence-only rescoring\", \"fast/full/max are ceilings not sizes\", \"Solar writes, Jev chooses NEXT ACTION\", \"SemIf 2186★ (+20 vs §109 2166)\", \"jevlike 1038★ (+7 vs 1031)\", \"TypeAR 14★ flat\", \"AnotiaWang 96★ (+1 vs 95)\", \"yibie/awesome-jev 490★\", \"Laya likes 802 (was 783)\", \"tracker likes 64 flat, lastModified UNCHANGED\", \"do not reopen or amend PR #23/#24/#25/#26/#27.\", \"do not reopen or amend PR #23 or #24 or #25 or #26 or #27\", \"Calibration is not alpha\", \"NO CURRENT ALPHA CANDIDATE\", \"ΔR² approximately +0.00084\", \"Brier 0.2131387\", \"ECE 0.0421875\", \"Adding Jev probability to deterministic volatility improved Brier by only 1.4058e-05\", \"default 0.5 keeps zero non pinned\", \"keepResult median 0.14 to 0.17\", \"keepCall median 0.28 to 0.35\", \"usable range is about 0.10 to 0.25\", \"7.8% to 57.9%\", \"judges results it never sees\", \"task-finish eval not built yet\", \"$0.002 per compaction\", \"slavadubrov/sgr-judge-bench ≠ slavadubrov/jev-judge-bench\", \"Jev 108/120 $0.083 0.34 s\", \"Luna SGR 114/120\", \"paired Jev accuracy-difference intervals include zero\", \"not evidence of equivalence\", \"GLM SGR 26/120 93 format failures\", \"Terra-planned Jev hybrid 55/120\", \"rule-based by default, optionally Jev-backed\", \"empty README\", \"missing key cannot break the experience\", \"prefill plus exactly one decode\", \"softmax over A/B/C ≠ Noul\", \"BBQ 9,053/10,000 (90.53%)\", \"ECE 0.0890\", \"Mean confidence 0.9943\", \"overconfident\", \"score and noul not implemented\", \"DGUI 12 rows (was 6)\", \"INSTRUCT 119 rows likes 2\", \"encode the state once, decide everything in parallel\", \"0.740 accuracy against a 0.508 majority\", \"ECE 0.047\", \"fine-tune's advantage ends where its 384-token training data does\", \"jasonkneen/open-jev ≠ pngwn/open-jev\", \"same sha d41dc3cd\", \"Space does not call Jev\", \"recomputes routing from saved probabilities\", \"200-case Jev 97.0% / 100.0% / 95.0% / MAE 9.22\", \"synthetic repository benchmark\", \"Jev evaluations are advisory\", \"YehuiTang0316/jev-nlgrep ≠ Bentlybro/jevgrep ≠ can1357/jegrep ≠ uehaj/jev-semgrep\", \"default threshold 0.8 still soft\", \"40-line windows cannot prove whole function\", \"token-native sequential start/end Choice\", \"Gemini/Haiku stubs not configured yet\", \"handful of hand-written examples, not a benchmark\", \"Jev judged exactly what it was given\", \"laguagu/jev-skills ≠ laguagu/jev-evidence-lab ≠ Pleo2/awesome-jev-agent-skills\", \"contract_passed is not a claim of guaranteed factual truth\", \"Wilson lower bound 0.85 floor\", \"fixture mode no savings claim\", \"SemIf 2207★ (+21 vs §110 2186)\", \"jevlike 1043★ (+5 vs 1038)\", \"TypeAR 15★ (+1 vs 14)\", \"AnotiaWang 97★ (+1 vs 96)\", \"yibie/awesome-jev 506★ (+16 vs 490)\", \"Laya likes 822 (was 802)\", \"tracker likes 64 flat, lastModified UNCHANGED\", \"do not reopen or amend PR #23/#24/#25/#26/#27/#28\", \"people who compare Jev against GPT-5.6 has never fine-tuned BERTForXYZ for living and it shows\", \"zero shot classifiers\", \"scale them as much as decoder only models\", \"many problems solved with LLMs could have been solved with them, it was a skill issue\", \"opt for DeBERTa and ModernBERT ones\", \"BERTForXYZ → DeBERTa → ModernBERT\", \"Jev vs GPT-5.6 bakeoffs are a category error\", \"encoder / ZS classifiers\", \"institutional HF voice\", \"quote *theirs*\", \"do not invent accuracy numbers\", \"softmax/ZS scores still ≠ calibrated Noul\", \"soft scores ≠ hard gates\", \"@mervenoyann\", \"likes 421 / 189\", \"impressions 35498 / 9613\", \"multimodal image<>text ZS as perception front-end\", \"hf:MoritzLaurer/deberta-v3-large-zeroshot-v2.0 likes 139\", \"hf:MoritzLaurer/ModernBERT-large-zeroshot-v2.0 likes 72\", \"Bart, bert, deberta, modernbert, these are all LLMs\", \"Maziyar quoted\", \"Jev is exemplar not the mandate\", \"do not reopen or amend PR #23/#24/#25/#26/#27/#28/#29\", \"soft Noul ≠ hard safety\" with a placement, not a stack replacement or a vendor how-to. Formal methods are one pillar. Not a substitute for the official typesafe-ai skill (live Jev API contracts)." license: MIT metadata: - version: 0.3.0 + version: 0.4.0 typesafe_skill: v0.5.7 typesafe_skill_commit: 65a39f3 tribute: "Named for Augustus De Morgan (1806-1871), mentor of William Stanley Jevons." diff --git a/.claude-plugin/marketplace.json b/.claude-plugin/marketplace.json index 6e15c3c1..b2aac596 100644 --- a/.claude-plugin/marketplace.json +++ b/.claude-plugin/marketplace.json @@ -4,10 +4,10 @@ }, "plugins": [ { - "description": "Design judgment for placing typed probabilistic judgment (Jev-class) using math, logic, and algorithmic mental models — across AI, SWE, business, knowledge work, and life. Not SWE-only. Formal methods are one pillar (Alloy vs Apalache; DST trio Antithesis/Resonate/PufferLib). GLiNER locate / GLiClass categorize are class peers, not footnotes. Expected utility, VOI, MCDA, signal detection, search/control, Leveson safety. Never launder a Noul as a proof. Not an API skill — load typesafe-ai for Jev contracts. Named for Augustus De Morgan, mentor of W. S. Jevons.", + "description": "Design judgment for placing typed probabilistic judgment (Jev-class System One) using math, logic, and algorithmic mental models — across AI, SWE, business, knowledge work, and life. Jev is the exemplar, not the monopoly (Laya, kev, OpenJev, TypeAR, GLiNER, encoder ZS). Ranking ≠ calibration; soft Noul ≠ hard gate. Formal methods are one pillar. Never launder a Noul as a proof. Not a TypeSafe product — load typesafe-ai for Jev contracts. Named for Augustus De Morgan, mentor of W. S. Jevons.", "name": "augustus", "source": "./.agents/skills/augustus", - "version": "0.3.0" + "version": "0.4.0" } ] } diff --git a/.github/workflows/scorecard.yml b/.github/workflows/scorecard.yml new file mode 100644 index 00000000..099735fa --- /dev/null +++ b/.github/workflows/scorecard.yml @@ -0,0 +1,49 @@ +# Official OpenSSF Scorecard action. +# https://github.com/ossf/scorecard-action +# Do not invent a numeric score in docs; the API is empty until this +# workflow publishes a result from default-branch runs. + +name: Scorecard supply-chain security +on: + branch_protection_rule: + schedule: + - cron: "30 1 * * 6" + push: + branches: ["main"] + +permissions: read-all + +jobs: + analysis: + name: Scorecard analysis + runs-on: ubuntu-latest + permissions: + security-events: write + id-token: write + contents: read + actions: read + + steps: + - name: Checkout code + uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1 + with: + persist-credentials: false + + - name: Run analysis + uses: ossf/scorecard-action@2d1146689b8cda280b9bc96326124645441f03bc # v2.4.4 + with: + results_file: results.sarif + results_format: sarif + publish_results: true + + - name: Upload artifact + uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1 + with: + name: SARIF file + path: results.sarif + retention-days: 5 + + - name: Upload to code-scanning + uses: github/codeql-action/upload-sarif@ff2f1c621b7f889edc0d3c761ac2e6a3f8cdb0dd # v4.37.7 + with: + sarif_file: results.sarif diff --git a/.gitignore b/.gitignore index e43b0f98..563ed30f 100644 --- a/.gitignore +++ b/.gitignore @@ -1 +1,17 @@ .DS_Store +.env +.env.* +!.env.example +*.pem +*.key +id_rsa +id_ed25519 +*.p12 +*.pfx +secrets/ +.secrets +credentials.json +*_credentials.json +.npmrc +.pypirc +*.sarif diff --git a/CHANGELOG.md b/CHANGELOG.md index 77f0f2fa..e01ae5e5 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -9,1371 +9,80 @@ revision it was written against. That skill owns integration contracts; Augustus owns design judgment. Re-read live TypeSafe docs before treating a pin as current API behavior. -## [0.3.0] - 2026-09-18 - -Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) -(`65a39f3`, 2026-09-12). - -### Added - -- User-provided HIGH Merve Noyan ZS classifier lineage - (`research/notes.md` §112): **Skip Archer rewrite.** - Docs-only on a **fresh PR off main**. - **HARD RULE:** do not reopen or amend PR #23 or #24 or - #25 or #26 or #27 or #28 or #29 (merged #29 owns - `notes.md` §111 / items 248–267 / batch #94 — leave it - alone). Institutional HF voice (@mervenoyann). Quote - *theirs*. Do not invent accuracy numbers. Jev vs - GPT-5.6 bakeoffs are a category error. encoder / ZS - classifiers (BERTForXYZ → DeBERTa → ModernBERT). many - problems solved with LLMs could have been solved with - them, it was a skill issue. opt for DeBERTa and - ModernBERT ones. multimodal image<>text ZS as - perception front-end. softmax/ZS scores still ≠ - calibrated Noul. soft scores ≠ hard gates. Live X this - pass: likes 421 / 189; impressions 35498 / 9613. - Hub likes 139 / 72 cited. X MCP used. - `invented_signal: false`. Composition items 268–272 / - batch #95. Uniqueness lock: people who compare Jev against GPT-5.6 has never fine-tuned BERTForXYZ for living and it shows; zero shot classifiers; scale them as much as decoder only models; many problems solved with LLMs could have been solved with them, it was a skill issue; opt for DeBERTa and ModernBERT ones; BERTForXYZ → DeBERTa → ModernBERT; Jev vs GPT-5.6 bakeoffs are a category error; encoder / ZS classifiers; institutional HF voice; quote *theirs*; do not invent accuracy numbers; softmax/ZS scores still ≠ calibrated Noul; soft scores ≠ hard gates; @mervenoyann; likes 421 / 189; impressions 35498 / 9613; multimodal image<>text ZS as perception front-end; hf:MoritzLaurer/deberta-v3-large-zeroshot-v2.0 likes 139; hf:MoritzLaurer/ModernBERT-large-zeroshot-v2.0 likes 72; Bart, bert, deberta, modernbert, these are all LLMs; Maziyar quoted; Jev is exemplar not the mandate; do not reopen or amend PR #23/#24/#25/#26/#27/#28/#29. +Hourly uniqueness locks from folds after v0.3.0 were moved out of this +file so release notes stay scannable. Archive: +[`research/changelog-hourly.md`](research/changelog-hourly.md). Canonical +folds: `research/notes.md`. -- Hourly 0646 HIGH (`research/notes.md` §111): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - **HARD RULE:** do not reopen or amend PR #23 or #24 or #25 - or #26 or #27 or #28 (do not reopen or amend PR #23/#24/#25/#26/#27/#28. - merged #23 owns `notes.md` §105 / items 149–160 / - batch #88; merged #24 owns `notes.md` §106 / items 161–177 / - batch #89; merged #25 owns `notes.md` §107 / items 178–185 / - batch #90; merged #26 owns `notes.md` §108 / items 186–201 / - batch #91; merged #27 owns `notes.md` §109 / items 202–225 / - batch #92; merged #28 owns `notes.md` §110 / items 226–247 / - batch #93 — leave them alone). Never reopen merged - #7–**#28**. Do **not** re-fold 0541 / §110 / 0439 / §109 / 0345 / §108 / 0243 / §107 / - 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / §103 / - 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 / §99 / - 1740 / §98 / 1639 / §96 / gliner-native-runtime / §97 / - 1541 / §95 / jev-align *mechanism* / §93 / - jev-orderby-bench *six-gates* / §60 / JevBench v1.2 - *board* / §78 / openJev-verdict *claim-audit* / §71 / - jev-judge-bench SLA-150 *contract* / §71 / - yuki-oshio/mini-jev *93.25%* / §103. How-to-apply / mental - models / architecture / Harbor-jevals / toolbelt — not a - thin Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating - ECE as alpha, 0.5 compaction as safety, 0.8 nlgrep as - 80% correctness, contract_passed as truth, Wilson 0.85 - as a proof, 90.53% BBQ as calibrated, softmax over - A/B/C as a Noul, 0.740 as TypeSafe vs Laya, 97.0% as - Harbor, or intervals-including-zero as equivalence is - soundness theater. Calibration is not alpha and ranking ≠ - calibration theater are the anti-patterns. Twenty HIGH - clusters / three themes: - **Measurement / benches / calibration PRIMARY** - ([alakise/calibration-is-not-alpha](https://github.com/alakise/calibration-is-not-alpha) - Calibration is not alpha; NO CURRENT ALPHA CANDIDATE; - ΔR² approximately +0.00084; Brier 0.2131387; ECE 0.0421875; - Adding Jev probability to deterministic volatility improved Brier by only 1.4058e-05; - Python MIT LICENSE SHA `cbfc2daf`; **0★**; HEAD `064b75f5`; README SHA `874ea57b`; size **268**); - ([OrMizL/jev-compaction-bench](https://github.com/OrMizL/jev-compaction-bench) - default 0.5 keeps zero non pinned; - keepResult median 0.14 to 0.17; keepCall median 0.28 to 0.35; - usable range is about 0.10 to 0.25; 7.8% to 57.9%; - judges results it never sees; task-finish eval not built yet; - $0.002 per compaction; - JavaScript MIT LICENSE SHA `3f4e7f2b`; **0★**; HEAD `92fd33e6`; README SHA `c7cd8b6e`; size **35**); - ([slavadubrov/sgr-judge-bench](https://github.com/slavadubrov/sgr-judge-bench) - slavadubrov/sgr-judge-bench ≠ slavadubrov/jev-judge-bench; - Jev 108/120 $0.083 0.34 s; Luna SGR 114/120; - paired Jev accuracy-difference intervals include zero; - not evidence of equivalence; - GLM SGR 26/120 93 format failures; Terra-planned Jev hybrid 55/120; - Python; SPDX NOASSERTION LICENSE SHA `690cf6b9`; **0★**; HEAD `5e142707`; README SHA `8026cafa`; size **0** WITH CONTENTS); - ([elyashium/atlas-replay-lab](https://github.com/elyashium/atlas-replay-lab) - rule-based by default, optionally Jev-backed; empty README; - missing key cannot break the experience; - JavaScript; GitHub license null; **0★**; HEAD `9856ab9b`; size **197**); - ([siren2345/jev-single-decode](https://github.com/siren2345/jev-single-decode) - prefill plus exactly one decode; softmax over A/B/C ≠ Noul; - BBQ 9,053/10,000 (90.53%); ECE 0.0890; Mean confidence 0.9943; - overconfident; score and noul not implemented; - Python MIT LICENSE SHA `3551844d`; **0★**; HEAD `65df86a3`; README SHA `51817d8c`; size **432**); - **datasets / Spaces** - (hfdataset:ctaxnagomi/DGUI_HYPERMEM-JEV - DGUI 12 rows (was 6); sha `ab3d3529`); - (hfdataset:ctaxnagomi/INSTRUCT_JEV - INSTRUCT 119 rows likes 2; sha `b0a09278`); - (hfspace:pngwn/open-jev - encode the state once, decide everything in parallel; - 0.740 accuracy against a 0.508 majority; ECE 0.047; - fine-tune's advantage ends where its 384-token training data does; - likes **25**; sha `d41dc3cd`); - (hfspace:jasonkneen/open-jev - jasonkneen/open-jev ≠ pngwn/open-jev; same sha d41dc3cd; likes **0**); - (hfspace:IkerMoel/open-alternative-jev densify sha `19b104c6`); - (hfspace:mobarmg/jev-schema-scorer densify sha `14d35d7e`; peaked ranking ≠ calibration); - (hfspace:reachjalil/jevlogs-triage-explorer - Space does not call Jev; recomputes routing from saved probabilities; - sha `dcb785ed`); - **applied / skills / economics** - ([IslamBaraka90/jev-typesafe-real-financial-use-cases](https://github.com/IslamBaraka90/jev-typesafe-real-financial-use-cases) - fifty recorded demos; recorded ≠ alpha; - JavaScript MIT LICENSE SHA `caaa17a8`; **0★**; HEAD `9d2eb48c`; README SHA `5329b92e`; size **6168**); - ([Milo318/mailordinal](https://github.com/Milo318/mailordinal) - 200-case Jev 97.0% / 100.0% / 95.0% / MAE 9.22; - synthetic repository benchmark; - TypeScript MIT LICENSE SHA `66faa16d`; **0★**; HEAD `19ea819d`; README SHA `a8df45fe`; size **109**); - ([Pleo2/awesome-jev-agent-skills](https://github.com/Pleo2/awesome-jev-agent-skills) - Jev evaluations are advisory; catalog ≠ endorsement; - Python MIT LICENSE SHA `25d9aa30`; **0★**; HEAD `42e3d179`; README SHA `efbe7ccf`; size **18**); - ([YehuiTang0316/jev-nlgrep](https://github.com/YehuiTang0316/jev-nlgrep) - YehuiTang0316/jev-nlgrep ≠ Bentlybro/jevgrep ≠ can1357/jegrep ≠ uehaj/jev-semgrep; - default threshold 0.8 still soft; 40-line windows cannot prove whole function; - TypeScript MIT LICENSE SHA `17cdbc7e`; **1★**; HEAD `ceec0d92`; README SHA `49e98f18`; size **6734**); - ([dangquan1402/jev-extract](https://github.com/dangquan1402/jev-extract) - token-native sequential start/end Choice; Gemini/Haiku stubs not configured yet; - Python MIT LICENSE SHA `d66f9c7a`; **0★**; HEAD `20c2f19f`; README SHA `9561e53d`; size **105**); - ([jyje/pilot-typesafeai-jev](https://github.com/jyje/pilot-typesafeai-jev) - handful of hand-written examples, not a benchmark; - Jev judged exactly what it was given; - Python MIT LICENSE SHA `08b2a315`; **0★**; HEAD `cabd4778`; README SHA `8d472372`; size **245**); - ([laguagu/jev-skills](https://github.com/laguagu/jev-skills) - laguagu/jev-skills ≠ laguagu/jev-evidence-lab ≠ Pleo2/awesome-jev-agent-skills; - MIT LICENSE SHA `e5ede121`; **0★**; HEAD `871ec586`; README SHA `a315944d`; size **72**); - ([lorensation/llm-cost-optimizer-jev](https://github.com/lorensation/llm-cost-optimizer-jev) - contract_passed is not a claim of guaranteed factual truth; - Wilson lower bound 0.85 floor; fixture mode no savings claim; - Python Apache-2.0 LICENSE SHA `261eeb9e`; **0★**; HEAD `db200b7e`; README SHA `6f2a595a`; size **227**). - SemIf 2207★ (+21 vs §110 2186); - jevlike 1043★ (+5 vs 1038); - TypeAR 15★ (+1 vs 14); - AnotiaWang 97★ (+1 vs 96); - yibie/awesome-jev 506★ (+16 vs 490); - Laya likes 822 (was 802); - tracker likes 64 flat, lastModified UNCHANGED; - Laya present; Blackwood ABSENT; Archer still promised_not_landed. - 0★ HIGH still got a real card. Soft Noul ≠ hard safety. - `invented_signal: false`. - Hourly 0646 uniqueness lock: Calibration is not alpha; NO CURRENT ALPHA CANDIDATE; ΔR² approximately +0.00084; Brier 0.2131387; ECE 0.0421875; Adding Jev probability to deterministic volatility improved Brier by only 1.4058e-05; default 0.5 keeps zero non pinned; keepResult median 0.14 to 0.17; keepCall median 0.28 to 0.35; usable range is about 0.10 to 0.25; 7.8% to 57.9%; judges results it never sees; task-finish eval not built yet; $0.002 per compaction; slavadubrov/sgr-judge-bench ≠ slavadubrov/jev-judge-bench; Jev 108/120 $0.083 0.34 s; Luna SGR 114/120; paired Jev accuracy-difference intervals include zero; not evidence of equivalence; GLM SGR 26/120 93 format failures; Terra-planned Jev hybrid 55/120; rule-based by default, optionally Jev-backed; empty README; missing key cannot break the experience; prefill plus exactly one decode; softmax over A/B/C ≠ Noul; BBQ 9,053/10,000 (90.53%); ECE 0.0890; Mean confidence 0.9943; overconfident; score and noul not implemented; DGUI 12 rows (was 6); INSTRUCT 119 rows likes 2; encode the state once, decide everything in parallel; 0.740 accuracy against a 0.508 majority; ECE 0.047; fine-tune's advantage ends where its 384-token training data does; jasonkneen/open-jev ≠ pngwn/open-jev; same sha d41dc3cd; Space does not call Jev; recomputes routing from saved probabilities; 200-case Jev 97.0% / 100.0% / 95.0% / MAE 9.22; synthetic repository benchmark; Jev evaluations are advisory; YehuiTang0316/jev-nlgrep ≠ Bentlybro/jevgrep ≠ can1357/jegrep ≠ uehaj/jev-semgrep; default threshold 0.8 still soft; 40-line windows cannot prove whole function; token-native sequential start/end Choice; Gemini/Haiku stubs not configured yet; handful of hand-written examples, not a benchmark; Jev judged exactly what it was given; laguagu/jev-skills ≠ laguagu/jev-evidence-lab ≠ Pleo2/awesome-jev-agent-skills; contract_passed is not a claim of guaranteed factual truth; Wilson lower bound 0.85 floor; fixture mode no savings claim; SemIf 2207★ (+21 vs §110 2186); jevlike 1043★ (+5 vs 1038); TypeAR 15★ (+1 vs 14); AnotiaWang 97★ (+1 vs 96); yibie/awesome-jev 506★ (+16 vs 490); Laya likes 822 (was 802); tracker likes 64 flat, lastModified UNCHANGED; do not reopen or amend PR #23/#24/#25/#26/#27/#28. +## [0.4.0] - 2026-09-20 -- Hourly 0541 HIGH (`research/notes.md` §110): **Skip - Archer rewrite.** Live-REST relock after adversarial - review (HEAD/README/size/stars/likes; Brier/Independent - design claims unchanged). Docs-only on a **fresh PR off main**. - **HARD RULE:** do not reopen or amend PR #23 or #24 or #25 - or #26 or #27 (do not reopen or amend PR #23/#24/#25/#26/#27. - merged #23 owns `notes.md` §105 / items 149–160 / - batch #88; merged #24 owns `notes.md` §106 / items 161–177 / - batch #89; merged #25 owns `notes.md` §107 / items 178–185 / - batch #90; merged #26 owns `notes.md` §108 / items 186–201 / - batch #91; merged #27 owns `notes.md` §109 / items 202–225 / - batch #92 — leave them alone). Never reopen merged - #7–**#27**. Do **not** re-fold 0439 / §109 / 0345 / §108 / 0243 / §107 / - 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / §103 / - 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 / §99 / - 1740 / §98 / 1639 / §96 / gliner-native-runtime / §97 / - 1541 / §95 / jev-align *mechanism* / §93 / - jev-orderby-bench *six-gates* / §60 / JevBench v1.2 - *board* / §78 / openJev-verdict *claim-audit* / §71 / - yuki-oshio/mini-jev *93.25%* / §103. How-to-apply / mental - models / architecture / Harbor-jevals / toolbelt — not a - thin Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating - Blackwood tracker-absent as landed, 0.9 as one number, - 8/8 as conversion lift, 200-row as ranking, 0.85 as 85%, - Hub Meanblock as a new species vs leesk212, circulating - $0.0004 as measured, Score as 0–1, Noul.confidence as - existing, json_schema gap as typed-model win, 64× Space - as Harbor, awesome listed counts as eval, WANLI 0.741 as - beating openjev, ECE 0.021 as a hard gate, Independent - `{cat,dog}` as Choice, softmax over letter slots as a - Noul, minProbability 0.85 as Harbor, or :max as a - mandatory search size is soundness theater. treating - 0.85 as 85% / minProbability hard-gate as Harbor, 0.9 - as one number, and 8/8 as conversion lift are the - anti-patterns. Twenty-two HIGH clusters / three themes: - **Open-weight / RLCD / Blackwood watch** - (hf:BlackwoodAI/blackwood-rlcd - Blackwood tracker ABSENT; likes 2 gated manual; - sha `3b9e29df`); - (hf:anthonym21/qwen3-0.6b-rlcd-decision - r = c - p_a; ECE 0.021; acc 0.807 vs warmup 0.746; - calibration beyond ~500 tokens unmeasured; - apache-2.0; likes **2**; sha `b327ec5e`); - (hf:larkooo/gemma-e2b-rlcd - Independent primitive; - 11.57s vs 54.10s · 4.67× · 120/128 *theirs*; - default path is pretrained Gemma probs not trained RLCD head; - apache-2.0; likes **1**; sha `e099c730`); - (hf:Meanblock/JEV-CPU - GH Meanblock 404; lock leesk212/JEV-CPU; - softmax over letter slots ≠ Noul; - mit; likes **7**; sha `759fa606`); - (hf:impacte/mimir-lfm-openjev - WANLI 0.741 vs openjev v2 0.77 *theirs*; - 3-way NLI ≠ Noul; likes **0**; sha `5f9173bb`); - (hf:shreyanbr/system-one-distilled | gold | zeroshot - priority 0.464 = majority floor; - banking77 contaminated; - raw margins not probabilities; - GH jev-haiku-benchmarking 404; - do not distill Jev as teacher of record (they distilled Haiku); - shas `56c9dba8` / `93e22fcf` / `4b6659d5`); - **measurement densifies PRIMARY** - ([Running-Dolphins/jev-bench](https://github.com/Running-Dolphins/jev-bench) - “0.9 is not one number”; ranking ≠ calibration; - banking77 0.8–0.9 stated 0.86 actual 0.73 over-confident *theirs*; - ≠ Praveenrajus/jev-bench ≠ fstandhartinger/jevbench; - Python MIT LICENSE SHA `891c70c6`; **0★**; HEAD `67d2ee42`; README SHA `ae2c5003`; size **1070**); - ([WallerChen/jev-measured](https://github.com/WallerChen/jev-measured) - $0.0000153–$0.0000226 vs circulating $0.0004 (~20×); - Score is 0..n-1 expectation not 0–1; - Noul has no confidence field; - TCP floor 198.8 ms; - type reliability is not a reason to choose Jev (json_schema 5/5); - gateway tax not one number; - Python MIT LICENSE SHA `0c563e25`; **0★**; HEAD `4a12dfb3`; README SHA `1c2c38ac`; size **37**); - ([RadRebelSam/jev-decision-lab](https://github.com/RadRebelSam/jev-decision-lab) - Function-only 5/8 vs hybrid 8/8; 4/8 without Jev; - 8 designed cases not conversion lift; - ≠ RadRebelSam/awesome-jev; - TS MIT LICENSE SHA `ae6f994c`; **0★**; HEAD `e6d6d42d`; README SHA `52a0071d`; size **128**); - ([jackojacko05/compare-jev-bigquery-ai-functions](https://github.com/jackojacko05/compare-jev-bigquery-ai-functions) - 200-row pilot Jev 86.5% 173/200 vs Gemini Flash-Lite 86.0% 172/200 vs Pro 87.0% 174/200 *theirs*; - not a ranking; - Jupyter MIT LICENSE SHA `0248c4b0`; **0★**; HEAD `cb9ef56b`; README SHA `d3db5d6c`; size **307**); - ([Trecto34/openjev-fighting-ring](https://github.com/Trecto34/openjev-fighting-ring) - NLI Tetris argmax P(entail)−P(contradict); - Python; license null; **0★**; HEAD `ac544f1e`; README SHA `b3fe2907`; size **1159**); - ([joshhu/jevtest](https://github.com/joshhu/jevtest) - 情緒測謊器; 1q 396ms / 30q 567ms; ±0.03; - 33q $0.000045 vs Gemini ~5× slower ~60× cost *theirs*; - ≠ realZachi/jevtest; - HTML; license null; **0★**; HEAD `6a4df41d`; README SHA `57d99325`; size **34**); - (hfspace:aahf/JevBenchmark - 8-example Jev vs GPT-5.6 Sol ~64× cost 5.4× latency *theirs*; - synthetic; no inference; ≠ JevBench v1.2 §78; - likes **1**; sha `36c28088`); - ([rhc98/awesome-jev](https://github.com/rhc98/awesome-jev) - Judged 3317 / listed 2560; Jev judges, code applies policy; - catalog ≠ endorsement; - TS; SPDX NOASSERTION LICENSE SHA `d491d714`; **1★**; HEAD `82898f25`; README SHA `e1f12080`; size **2529**); - ([AiPersonacademy/Awesome-jev-use](https://github.com/AiPersonacademy/Awesome-jev-use) - APA “microsecond policy / zero hallucination” overclaim; - CC0-1.0 LICENSE SHA `7ba0a23e`; **2★**; HEAD `29246f12`; README SHA `d506c11f`; size **124**); - **applied / theory placements** - ([erseco/questionator](https://github.com/erseco/questionator) - Client-side quiz; pointer from held docs; scanned-PDF warn; - CSP only api.typesafe.ai; - JS MIT LICENSE SHA `3ebdd11d`; **0★**; HEAD `b498f10b`; README SHA `4334c847`; size **1267**); - ([grgy078033/grill-jev](https://github.com/grgy078033/grill-jev) - Jev judges / agent reasons / user decides; - selecting an option is not permission to implement; - degraded fallback; - Python MIT LICENSE SHA `7c0e962d`; **1★**; HEAD `5ce6cbe8`; README SHA `da80efcc`; size **85**); - ([makefunstuff/jev-lsp](https://github.com/makefunstuff/jev-lsp) - pattern exact, judgement must clear floor; - no matching pattern → no model call; - not a correctness oracle; - $0.00022 vs chat $0.00306 *theirs*; - Rust MIT LICENSE SHA `fb5bc4a5`; **0★**; HEAD `bce3d8ed`; README SHA `c7ecbf8f`; size **1696**); - ([nozomi-koborinai/jev-spec](https://github.com/nozomi-koborinai/jev-spec) - Spec vs artifact remainder; - treating 0.85 as 85% / minProbability hard-gate as Harbor; - TS MIT LICENSE SHA `e43f732a`; **1★**; HEAD `86f14198`; README SHA `c0f29baa`; size **73**); - ([202620325-spec/Jev-LLM](https://github.com/202620325-spec/Jev-LLM) - VERIFY acquires discriminating evidence, never same-pool confidence-only rescoring; - fast/full/max are ceilings not sizes; - Solar writes, Jev chooses NEXT ACTION; - Python MIT LICENSE SHA `5d699c97`; **0★**; HEAD `da06b6d1`; README SHA `f371ba83`; size **150**). - SemIf 2186★ (+20 vs §109 2166); - jevlike 1038★ (+7 vs 1031); - TypeAR 14★ flat; - AnotiaWang 96★ (+1 vs 95); - yibie/awesome-jev 490★; - Laya likes 802 (was 783); - tracker likes 64 flat, lastModified UNCHANGED; - Laya present; Blackwood ABSENT; Archer still promised_not_landed. - 0★ HIGH still got a real card. Soft Noul ≠ hard safety. - `invented_signal: false`. - -- Hourly 0439 HIGH (`research/notes.md` §109): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - **HARD RULE:** do not reopen or amend PR #23 or #24 or #25 - or #26 (do not reopen or amend PR #23/#24/#25/#26. - merged #23 owns `notes.md` §105 / items 149–160 / - batch #88; merged #24 owns `notes.md` §106 / items 161–177 / - batch #89; merged #25 owns `notes.md` §107 / items 178–185 / - batch #90; merged #26 owns `notes.md` §108 / items 186–201 / - batch #91 — leave them alone). Never reopen merged - #7–**#26**. Do **not** re-fold 0345 / §108 / 0243 / §107 / - 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / §103 / - 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 / §99 / - 1740 / §98 / 1639 / §96 / gliner-native-runtime / §97 / - 1541 / §95 / jev-align *mechanism* / §93 / - jev-orderby-bench *six-gates* / §60 / JevBench v1.2 - *board* / §78 / openJev-verdict *claim-audit* / §71 / - yuki-oshio/mini-jev *93.25%* / §103. How-to-apply / mental - models / architecture / Harbor-jevals / toolbelt — not a - thin Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating - WANLI-256 64.5/60.2/52.0 as Harbor, label_mass as - correctness, Hub v1 family cards as shipping GGUF, - re-folding §108 27B/ternary as new, treating LoRA stub - cards as independent eval, hard-gating jevify n=307 ECE - 0.061 as “honest probabilities”, collapsing kushalpatil - into Mintzs/gulagala001/uspraveen, treating GH - kushalpatil07/jevify as an existing repo, collapsing - JulesHuisman into SivletLabs/4esv, quoting 32/32 synthetic - as production, hard-gating 0.8 evidence as proof, quoting - 40-band 0/10 without 7-band, treating /judge 0.5 as truth, - treating git-confess 11% as a person verdict, treating - paper-trader +12.40% as edge, treating Awesomejev - 656/38160 as eval, treating tracker likes as Archer - landing, using roadus2 spelling, pasting ultra_laya vs-Jev - as this-fork win, or treating jevgraph 100% gated as - production is soundness theater. jevify ECE 0.061 as a - hard gate, 0.8 evidence as proof, and paper-trader fills - as edge are the anti-patterns. Twenty-four HIGH clusters / - three themes: - **open reproduction densifies** - (hf:kushalpatil/jevify-gemma4-26b-a4b - Gemma-4 26B-A4B jevify classification+calibration; - Hub jevify merged LoRA ships weights; - PAWS 0.580/ece 0.288 is the weak cell; - kushalpatil/jevify-gemma4 ≠ Mintzs/jevify ≠ gulagala001/jevify ≠ uspraveen/Jevify; - GH kushalpatil07/jevify 404; - license gemma; likes **0**; sha `d4c0d1d4`); - (hf:kushalpatil/jevify-gemma4-26b-a4b-lora - LoRA adapter twin not independent eval; - license null; likes **0**; sha `ec4a3d22`); - (hf:kushalpatil/jevify-gemma4-e4b - Gemma-4 E4B jevify; - smaller E4B slightly better OOD ECE than 26B-A4B; - license gemma; likes **0**; sha `a6b5a716`); - (hf:kushalpatil/jevify-gemma4-e4b-lora - E4B LoRA stub card; license null; likes **0**; sha `cca1f55e`); - (hf:NicolaiMTLassen/bonzi-8b-v1-jev - bonzi Bonsai-8B v1 GGUF densify; WANLI-256 64.5% *theirs*; - rank #4 of 6; MIT; likes **0**; sha `588bc44e`); - (hf:NicolaiMTLassen/bonzi-1.7b-v1-jev - Bonsai-1.7B v1; WANLI-256 52.0% *theirs*; rank #6 of 6; - MIT; likes **0**; sha `48148bf9`); - (hf:NicolaiMTLassen/bonzi-4b-v1-jev - Bonsai-4B v1; WANLI-256 60.2% *theirs*; rank #5 of 6; - MIT; likes **0**; sha `d5545084`); - ([roadius2/ultra_laya](https://github.com/roadius2/ultra_laya) - roadus2 watch misspelling; lock roadius2/ultra_laya; - ultra_laya REVIEW defects; - default branch claude/laya-jev-review-gg5ppo; - Python Apache-2.0 LICENSE SHA `d9a10c0d`; **0★**; HEAD `0dff5bd2`; README SHA `da95bc05`; size **0** WITH CONTENTS); - **measurement densifies PRIMARY** - ([AHTOOOXA/jev-cyrillic-audit](https://github.com/AHTOOOXA/jev-cyrillic-audit) - XNLI EN 88.3% ECE 0.032 → RU 77.3% ECE 0.096; - Δ −11.0 pp [−14.2,−7.8]; ECE +0.063; - MASSIVE no detectable difference at n=600; - confidence is function of p_max (r=1.000); - Python MIT LICENSE SHA `f9b9193e`; **0★**; HEAD `7167894a`; README SHA `d36a36f5`; size **1302**); - ([JulesHuisman/jev-eval](https://github.com/JulesHuisman/jev-eval) - JulesHuisman/jev-eval scaffolding / README SHA c356a584 (was empty e69de29b); - JulesHuisman/jev-eval ≠ SivletLabs/jev-eval ≠ willkelly/jev-evaluation ≠ 4esv ≠ xxkuboxx ≠ onlyoneaman ≠ dayhaysoos/jevals; - Python; license null; **0★**; HEAD `96c2a110`; size **0** WITH CONTENTS); - ([tunahansahin897/what-is-jev](https://github.com/tunahansahin897/what-is-jev) - 947 repos scored; A 273 / B 302 / C 372; LLM rubric ≠ benches; - Python; SPDX NOASSERTION LICENSE SHA `790f3358`; **0★**; HEAD `71d53be2`; README SHA `3ae98c56`; size **6277**); - ([Mishkun/judge-jev](https://github.com/Mishkun/judge-jev) - judge-jev 0.5 still soft; - TS; license null; **0★**; HEAD `1bf495d3`; README SHA `3788ab6e`; size **65**); - **applied class placements** - ([LiuHao-1443/jev-table-tennis](https://github.com/LiuHao-1443/jev-table-tennis) - 7 bands 6/10 vs 40 bands 0/10; - Python MIT LICENSE SHA `777dc3d6`; **1★**; HEAD `0224c21c`; README SHA `3a54bc93`; size **0** WITH CONTENTS); - ([laguagu/jev-evidence-lab](https://github.com/laguagu/jev-evidence-lab) - source receipts + confidence slider re-policy without re-inference; - 32/32 synthetic is smoke not production; - Python MIT LICENSE SHA `e5ede121`; **0★**; HEAD `a9669e94`; README SHA `1be53f55`; size **0** WITH CONTENTS); - ([hemanth/hfjev](https://github.com/hemanth/hfjev) - classify HF datasets across typed semantic dimensions; - Python MIT LICENSE SHA `0ac06d77`; **1★**; HEAD `6f2aa501`; README SHA `20651a25`; size **48**); - ([akash-kamat/jev-llm](https://github.com/akash-kamat/jev-llm) - pointer-not-generator 400 human-authored responses; - JS; license null; **0★**; HEAD `194db1a6`; README SHA `c42907a6`; size **54**); - ([chenmingtang830/jevgraph](https://github.com/chenmingtang830/jevgraph) - proposed ≠ authorized; FewRel 160: Jev 85.0% vs lexical 13.125%; - gated 100% (95/95) coverage 59.375%; - Python Apache-2.0 LICENSE SHA `57bc88a1`; **0★**; HEAD `7a6f7d05`; README SHA `bda2ed37`; size **125**); - ([Towow-ai/jpp](https://github.com/Towow-ai/jpp) - J++ composable semantic computation language; - Python MIT LICENSE SHA `b1a46fbc`; **3★**; HEAD `14d77789`; README SHA `0a82d97f`; size **1071**); - ([whyashthakker/awesome-jev-use-cases](https://github.com/whyashthakker/awesome-jev-use-cases) - No benchmark winner is claimed; - HTML MIT LICENSE SHA `13b5a2a6`; **3★**; HEAD `74583663`; README SHA `a6d96b87`; size **426**); - ([dog-last/awesome-jev](https://github.com/dog-last/awesome-jev) - phishing: naive 62.6% vs regex 91.8%; 5-atomic + LR 95.0% *theirs*; - Python MIT LICENSE SHA `cd31b9aa`; **1★**; HEAD `206fdcab`; README SHA `1ca65f69`; size **85**); - ([shinshin86/jev-aituber-tension-sample](https://github.com/shinshin86/jev-aituber-tension-sample) - AITuber tension ±15; - TS MIT LICENSE SHA `8b20d89e`; **0★**; HEAD `8c0a8ffb`; README SHA `06536d19`; size **105**); - ([shinpr/jev-reranker](https://github.com/shinpr/jev-reranker) - README npm global; repo is Rust; - Rust MIT LICENSE SHA `4306a712`; **1★**; HEAD `731deba3`; README SHA `e2818c0e`; size **163**); - ([AHTOOOXA/git-confess](https://github.com/AHTOOOXA/git-confess) - git-confess code owns counting/blame/ratio; - httpx exhibit 11% (13/119) *theirs*; - Python MIT LICENSE SHA `f9b9193e`; **0★**; HEAD `54cd2849`; README SHA `764076a6`; size **3**); - ([waterme7on/jev-paper-trader](https://github.com/waterme7on/jev-paper-trader) - 90d trend +12.40% vs random +12.75% vs BH +41.71%; - 5m win rate 25%; - JS; license null; **0★**; HEAD `73662f79`; README SHA `4eded35a`; size **145**). - Awesomejev 656 entries / 38,160 stars; - WANLI-256 64.5% / 60.2% / 52.0% *theirs*; - rank #4 / #5 / #6 of 6; - tracker likes 64 (+4) lastModified UNCHANGED; - Laya present; Blackwood ABSENT; Archer still promised_not_landed. - 0★ HIGH still got a real card. Soft Noul ≠ hard safety. - `invented_signal: false`. +Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) +(`65a39f3`, 2026-09-12). Live HEAD of that repo is still this commit — +the only tagged official-skill revision. -- Hourly 0345 HIGH (`research/notes.md` §108): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - **HARD RULE:** do not reopen or amend PR #23 or #24 or #25 (merged - #23 owns `notes.md` §105 / items 149–160 / batch #88; - merged #24 owns `notes.md` §106 / items 161–177 / batch #89; - merged #25 owns `notes.md` §107 / items 178–185 / batch #90 — - leave them alone). Never reopen merged - #7–**#25**. Do **not** re-fold 0243 / §107 / 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / - §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 - / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime - / §97 / 1541 / §95 / jev-align *mechanism* / §93 / - jev-orderby-bench *six-gates* / §60 / JevBench v1.2 - *board* / §78 / openJev-verdict *claim-audit* / §71 / - yuki-oshio/mini-jev *93.25%* / §103. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating - WANLI-256 as Harbor, label_mass as correctness, Hub - family cards as shipping weights, collapsing stock Q1_0 - into §107 Q2_0 gibberish, treating the HF twin as a - weights drop, quoting 93.25% as family-disjoint, - hard-gating injection ECE 0.058 as “Jev is calibrated”, - quoting AbstentionBench rank 1 without question-asymmetry / - 2025 field, treating dspy-bench as a quality claim, - hard-gating pdf-race 12/12 as pipeline equality, treating - atlas listed counts as eval, hard-gating flopcheck - composite as truth, treating calibration-lab 40-row T as - production, treating CLEVR joint 0% as “vision Jev works”, - treating ONNX 63/63 as ECE, treating openkev T as - transferable, or treating select_threshold inf as a bug - is soundness theater. Injection ECE 0.058 as a hard gate - and 40-row T as production are the anti-patterns. Sixteen - HIGH clusters / two themes: - **open reproduction densifies** - (hf:NicolaiMTLassen/bonzi-27b-v2-jev - ternary Bonsai 27B v2 open-bonzi GGUF family densify; - Hub still does not ship weights; WANLI-256 74.6% *theirs*; - ternary still needs PrismML fork; MIT; likes **0**; sha `c4d21b74`); - (hf:NicolaiMTLassen/bonzi-8b-ternary-v1-jev - ternary Bonsai 8B open-bonzi GGUF; WANLI-256 65.2% *theirs*; - MIT; likes **0**; sha `47b66187`); - (hf:NicolaiMTLassen/bonzi-27b-v1-jev - Bonsai 27B v1 open-bonzi GGUF; Bonsai 1 27B Q1_0 runs on stock llama.cpp; - WANLI-256 71.1% *theirs*; MIT; likes **0**; sha `2fb8061a`); - (hf:mizchi/laya-multilingual-onnx - Laya multilingual ONNX WebGPU typed-decisions port; - 63/63 selected answers / 5.1e-4 CPU / 1.2e-2 WebGPU; - ships model.onnx 646.9MB; apache-2.0; likes **0**; sha `b6314ec9`); - (hf:IamBusy/OpenJev-Vision - OpenJev Vision image classification + uncertainty; - CLEVR-4 held-out joint 0%; license other; likes **0**; sha `8cf6cbd3`); - (hf:heman10x/openJev-verdict-2.0 - twin tokenizer-only; no 149.6M weights; likes **5** ≠ GH **107★**; - apache-2.0; sha `794d5e0d`); - (hfdataset:IamBusy/OpenJev-Vision-Research-v0.1 - 12,832 images; 294,912 derived targets not independent samples; - license other; likes **0**; sha `43e49184`); - ([UpHash-Network/mini-jev](https://github.com/UpHash-Network/mini-jev) - UpHash-Network/mini-jev is yuki-oshio transfer; - residual-head 9,222-param decreased 73/96→67/96; - Python MIT LICENSE SHA `79e2cff7`; **0★**; HEAD `52fbae12`; README SHA `363441b6`; size **38388**); - **measurement densifies PRIMARY** - ([ASEVlad/jev-injection-bench](https://github.com/ASEVlad/jev-injection-bench) - jev-injection-bench 11,900 labelled prompts; - Jev best ranking / Haiku better ECE 0.021 vs 0.058; - 0.5–0.9 band is where Jev's numbers do not mean what they say; - Prompt wording moves panic 28%; - Python MIT LICENSE SHA `0dab34c6`; **0★**; HEAD `c0d0f25d`; README SHA `1b28498c`; size **107**); - ([manojlds/jev-dspy-bench](https://github.com/manojlds/jev-dspy-bench) - Jev agreement is similarity, never ground truth; - no aggregate quality grade or merge gate; - manojlds/jev-dspy-bench ≠ dspachos/jev-dspy ≠ jmanhype/jev-dspy-lab; - Python Apache-2.0 LICENSE SHA `261eeb9e`; **0★**; HEAD `d8c68d72`; README SHA `16100deb`; size **0** WITH CONTENTS); - ([sshariqali/jev-abstentionbench](https://github.com/sshariqali/jev-abstentionbench) - AbstentionBench-on-Jev rank 1 of 20 vs 2025 field; - question-asymmetry; forward-looking 0.465 never extreme; - Python; README MIT / GitHub SPDX NOASSERTION LICENSE SHA `7f6fb3ba`; **0★**; HEAD `f5c0c068`; README SHA `26393f0f`; size **0** WITH CONTENTS); - ([misakaikato/openkev](https://github.com/misakaikato/openkev) - openkev calibration layer not a runtime; - ECE vs coverage independent; select_threshold returns inf; - escalation catches uncertainty not ignorance; - misakaikato/openkev ≠ jaredpalmer/kev; - Python MIT LICENSE SHA `8c3231cb`; **0★**; HEAD `babcab1c`; README SHA `974d97f9`; size **484**); - ([goodrahstar/pdf-race](https://github.com/goodrahstar/pdf-race) - pdf-race Docling→Jev vs Gemini; parser owns the wall clock; - 12/12 tie is a tie; titles selected not generated; - JS MIT LICENSE SHA `182ac1d5`; **0★**; HEAD `1c687fc6`; README SHA `05730787`; size **0** WITH CONTENTS); - ([ZeroX-01/jev-atlas](https://github.com/ZeroX-01/jev-atlas) - ZeroX-01/jev-atlas ≠ Zaious/jev-capability-atlas ≠ gorock007/jev-atlas; - catalog not endorsement; - JS; license null; **0★**; HEAD `bc94bf50`; README SHA `8dd3d308`; size **0** WITH CONTENTS); - ([samyakjain0606/jev-is-here](https://github.com/samyakjain0606/jev-is-here) - flopcheck 16 calibrated tweet judgments; mechanical tells in code; - TS; license null; **0★**; HEAD `eb0de0ba`; README SHA `cefb8f96`; size **0** WITH CONTENTS); - (hfspace:BunsDev/laya-calibration-lab - Laya calibration lab Gradio MCP; T never changes argmax; - confidence ≠ top-label p; easy probe set refused; - 40–48 rows too small to ship T; apache-2.0; likes **0**; sha `a3fc13ba`). - 0★ HIGH still got a real card. Soft Noul ≠ hard safety. - `invented_signal: false`. +Twenty-eight commits on `main` after the v0.3.0 tag (merged #2–#30, +through Merve §112). Open #31 (0743 fold) is in flight on another +branch and is not part of this release. -- Hourly 0243 HIGH (`research/notes.md` §107): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - **HARD RULE:** do not reopen or amend PR #23 or #24 (merged - #23 owns `notes.md` §105 / items 149–160 / batch #88; - merged #24 owns `notes.md` §106 / items 161–177 / batch #89 — - leave them alone). Never reopen merged - #7–**#24**. Do **not** re-fold 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / - §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 - / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime - / §97 / 1541 / §95 / jev-align *mechanism* / §93 / - jev-orderby-bench *six-gates* / §60 / JevBench v1.2 - *board* / §78 / openJev-verdict *claim-audit* / §71. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating a - 0.85 figure FAST_PATH as a proof, 100/100 easy T/F as - Harbor, label_mass as correctness, 77.10% as beating - Jev, collapsing ticket-router into Sarath, collapsing - jev-bench into jevbench, collapsing DeBERTa ONNX into - the Qwen scorer ONNX, collapsing Hub/GH Nicolai - spellings, treating 151M vs 149.6M as two models, - quoting 62.3% cost save without the 4.5pp miss, - treating mock keyword as production, hard-gating - ticket 0.6 as safety, treating a study mock as a live - API, treating stock llama.cpp Q2_0 as working, or - re-folding §71 as a beat is soundness theater. - Figure-router 0.85 hard-gate and 100/100 easy T/F as - Harbor are the anti-patterns. Eight HIGH clusters / - three themes: - **measurement densifies PRIMARY** - ([erendikmenn/jev-llm-router-benchmark](https://github.com/erendikmenn/jev-llm-router-benchmark) - Benchmark-driven Jev router and judge; cheap alone is not success; - Jev does not write, sum prices, or claim accuracy %; - Sol 94.2 / Luna 83.9 / Jev path 89.7; - 19.2% Sol / 62.3% cost save / 4.5pp miss of 2pp non-inferiority; - p50 latency worse than Sol due to routing overhead; - erendikmenn/jev-llm-router-benchmark ≠ jev-rag-benchmark ≠ ryantsai/jev-llm-router; - Python MIT LICENSE SHA `99916677`; **0★**; HEAD `f44ef450`; README SHA `df3687e5`; size **0** with contents); - ([aesaganda/jev-ticket-router](https://github.com/aesaganda/jev-ticket-router) - Express + node:sqlite; mock and Jev decision engines; - previous_ticket_count >= 3 is code; MIN_CONFIDENCE 0.6 still soft; - substring false positives; - aesaganda/jev-ticket-router ≠ SarathChandraBellam/jev-vs-llm-ticket-router; - JS; license null; **0★**; HEAD `ecf00046`; README SHA `41c0b50f`; size **0** with contents); - ([hoangngochuong24947-gif/jev-figure-router](https://github.com/hoangngochuong24947-gif/jev-figure-router) - Universal Figure & Diagram Router; confidence ≥ 0.85 hard-gate is theater; - generative AI banned from scientific plots; six visual branches; - Python MIT LICENSE SHA `62120c23`; **1★**; HEAD `4c51b1af`; README SHA `11d5302b`; size **340**); - (hfdataset:Praveenrajus/jev-bench - human-labeled (state, question, label); 166,054 rows / 22 configs; - soft_label for human uncertainty; - Praveenrajus/jev-bench ≠ fstandhartinger/jevbench; - license other; likes **0**; sha `c9c3032c`); - **open reproduction class ports** - (hf:NicolaiMTLassen/open-bonzi-jev - ternary bonsai System One GGUF; openjev's mechanism, Bonsai's weights; - Hub does not ship weights; 100/100 easy T/F is not Harbor; - label_mass ≠ correctness; stock llama.cpp Q2_0 silently gibberish; - NicolaiMTLassen/open-bonzi-jev ≠ NicolaiLassen; - MIT; likes **0**; sha `09240156`; - GH companion NicolaiLassen/open-bonzi-jev Python MIT; **0★**; HEAD `1c2508fd`; LICENSE SHA `46d740e4`; size **0** with contents); - (hf:onnx-community/open-jev-deberta-v3-large-ONNX - transformers.js DeBERTa ONNX; source:com-kotobalabs/open-jev-deberta-v3-large; - temperature 1.05; AutoModel from_pretrained works; - onnx-community/open-jev-deberta-v3-large-ONNX ≠ system-one-qwen3.5-4b-scorer-ONNX; - apache-2.0; likes **0**; sha `3bc2553b`); - ([Heman10x-NGU/openJev-verdict-2.0](https://github.com/Heman10x-NGU/openJev-verdict-2.0) - 107★ densify; GH 151M vs README 149.6M; PR #1 now closed unmerged; - do not re-fold §71 claim-audit as a beat; - Python; README Apache-2.0 / GitHub SPDX NOASSERTION; **107★**; HEAD `a458733c`; README SHA `05ca75af`; size **14728**); - **study densification** - ([wjdjdakf17/jev-study](https://github.com/wjdjdakf17/jev-study) - typed decisions, RLCD, confidence-gated routing; structured ≠ correct; - mock not live API; 26 tests; wjdjdakf17/jev-study ≠ baekenough/jev-study; - TypeScript MIT LICENSE SHA `00bbf5f9`; **0★**; HEAD `24b5d7d7`; README SHA `4ca30c93`; size **34**; default **master**). - 0★ HIGH still got a real card. Soft Noul ≠ hard safety. - `invented_signal: false`. +### Added -- Hourly 0145 HIGH (`research/notes.md` §106): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - **HARD RULE:** do not reopen or amend PR #23 (merged - #23 owns `notes.md` §105 / items 149–160 / batch #88 — - leave it alone). Never reopen merged - #7–**#23**. Do **not** re-fold 0042 / §105 / 2340 / §104 / 2246 / - §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 - / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime - / §97 / 1541 / §95 / jev-align *mechanism* / §93 / - jev-orderby-bench *six-gates* / §60 / JevBench v1.2 - *board* / §78. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating a - Jevify two-liner as a checkpoint, collapsing Jevify - into Mintzs or gulagala001, pasting Colvin as hyusi, - treating JSON parse as a Noul, quoting 46x / 91.1% / - 92.58% / 84.8 as class ceilings, treating - classifier.dev #1 as a better model, re-folding §78 - as new, letting Jev do budget arithmetic, treating - XState as Jev, pasting catalog ★ as eval, treating - SPLADE as TypeSafe Jev, or treating an empty - consistency Space as a win is soundness theater. - Jevify stub and JSON-parse-as-Noul are the - anti-patterns. Seventeen HIGH clusters / four themes: - **architecture probes PRIMARY** - ([uspraveen/Jevify](https://github.com/uspraveen/Jevify) - Turn any open LLM into System-One Jev; - uspraveen/Jevify ≠ Mintzs/jevify ≠ gulagala001/jevify; - Jevify-any-LLM architecture probe; description-only stub / size 0; - license null; **0★**; HEAD `0f29d783`; README SHA `32608473`); - ([Ruivalim/exu-base](https://github.com/Ruivalim/exu-base) - Train encoder-only calibrated decision models from a task sentence; - Exu is a toolkit, not a method; strictly proper scoring rule; Pre-alpha; - Python MIT LICENSE SHA `336cde4c`; **1★**; HEAD `7288cdca`; README SHA `45244838`; size **180**); - ([Colvin0315/MiniSystemOne](https://github.com/Colvin0315/MiniSystemOne) - scratch-trained calibrated decision model; typed Q → probability dists; - Colvin0315/MiniSystemOne ≠ hyusi2003/MiniSystemOne; - no published weights download URL; 90.5 seconds / 29.2% pipeline evidence; - p_i/p_j independent of other candidates; - Apache-2.0 LICENSE SHA `261eeb9e`; **0★**; HEAD `d7f9f803`; README SHA `a5b0d2fd`; size **814**); - ([scienthoon/luce](https://github.com/scienthoon/luce) - Recipe for calibrated decision models — small model out; - init → synth → train → eval → serve; 91.1 % / ECE 0.022 *theirs*; - Jev zero-shot 75.1; Apache-2.0 LICENSE SHA `d6456956`; **1★**; HEAD `8072b97d`; README SHA `5fbe226d`; size **10833**); - ([RichardoMrMu/jev-mini](https://github.com/RichardoMrMu/jev-mini) - Put Jev's three headline claims on trial; 0.5B local GPU; - 46x speedup / accuracy identical; ECE 0.624 sentiment catastrophe; - bigger model worse calibration; - RichardoMrMu/jev-mini ≠ yuki-oshio/mini-jev ≠ r-ms/mini-jev; - Python MIT LICENSE SHA `cbdf24bb`; **0★**; HEAD `5adef5dc`; README SHA `154ae8b0`; size **0** with contents); - ([tapsin/jev-local](https://github.com/tapsin/jev-local) - System-1 decision engine for local LLMs; structured choices only; - JSON parse of generated text ≠ Noul; TypefAI JEV / Journal Entry Voucher; - tapsin/jev-local ≠ us/jev-local ≠ Argos1111/jev_local; - Python; license null; **0★**; HEAD `96aac2a1`; README SHA `4d728cbf`; default **master**; size **0** with contents); - **measurement densifies** - ([goya4140/jev-reward-model-evaluation](https://github.com/goya4140/jev-reward-model-evaluation) - Jev 1.13 reward-model eval across 8 benchmark tracks; - 40,940 examples / 0 API errors; RewardBench v1 92.58%; Precise IF 50.63%; - Python MIT LICENSE SHA `4588ffe5`; **0★**; HEAD `f16a06d1`; README SHA `603587fd`; size **0** with contents); - ([SarathChandraBellam/jev-vs-llm-ticket-router](https://github.com/SarathChandraBellam/jev-vs-llm-ticket-router) - Jev vs LLM support-ticket routing; Scaffolding in progress; - license null; **0★**; HEAD `0318ad72`; README SHA `8223dd6f`; default **master**; size **0**); - ([Shilin237/jev-vs-llm-cost](https://github.com/Shilin237/jev-vs-llm-cost) - static + live decision bench; TypeSafe's own published benchmark; - illustrative simulations, not live API calls; - HTML; license null; **0★**; HEAD `2d4bd6a9`; README SHA `f4e98bbf`; default **master**; size **0**); - ([fstandhartinger/jevbench](https://github.com/fstandhartinger/jevbench) - JevBench v1 — smart/cheap/fast/reliable; I/C/S/K 25% geometric mean; - classifier.dev fast tier 84.8 is Jev behind its own API; - do not re-fold §78 v1.2 board as new; Laya (421M) 70.1 now on board; - Python MIT LICENSE SHA `ebdd5738`; **6★**; HEAD `c7ab99f5`; README SHA `8fe07c41`; size **8907**); - ([AIGNLAI/ReflexRoute](https://github.com/AIGNLAI/ReflexRoute) - Zero-shot/few-shot LLM routing; hard budget filter before Jev; - Jev never asked to perform budget arithmetic; - Python MIT LICENSE SHA `1be8631e`; **1★**; HEAD `5f475d80`; README SHA `6a1ae694`; size **0** with contents); - ([priyankark/jev-state](https://github.com/priyankark/jev-state) - Jev judges the next state, XState enforces transitions; - simulation uses synthetic keyword fixtures; - TypeScript MIT LICENSE SHA `78745387`; **0★**; HEAD `c73aef65`; README SHA `15410dbe`; size **0** with contents); - (hfspace:mjyoke1111/jev-consistency-benchmark - Consistency benchmark Space; This Space contains no benchmark result yet; - 12-case plumbing fixture; apache-2.0; likes **0**; lastModified `2026-09-20T05:29:45Z`); - **catalog gravity** - ([v-modal/awesome-jev-tools](https://github.com/v-modal/awesome-jev-tools) - catalog gravity; ★339 live REST; curation is not endorsement; - license null; **339★**; HEAD `f117e0c3`; README SHA `8ea9a669`; size **83**); - ([RadRebelSam/awesome-jev](https://github.com/RadRebelSam/awesome-jev) - crawler-maintained directory; Daily GitHub + npm sweep, human-merged; - RadRebelSam/awesome-jev ≠ AnotiaWang ≠ yibie ≠ cobanov ≠ logicrw ≠ v-modal; - SPDX NOASSERTION LICENSE SHA `2aa7fe23`; **0★**; HEAD `27629954`; README SHA `21ec8d51`; size **919**); - **HF class ports** - (hf:rdxtremity/jev-reranking - HF peft SPLADE/BGE reranker; rdxtremity/jev-reranking ≠ carlaiau/jev-reranking; - query-side encoders, not a Jev replica; apache-2.0; likes **0**; sha `cae796ea`); - (hf:onnx-community/system-one-qwen3.5-4b-scorer-ONNX - ONNX System One Qwen3.5-4B scorer; source:pngwn/system-one-qwen3.5-4b-scorer; - CC-BY-NC-4.0; temperature 1.75; transformers.js AutoModel cannot load this graph; - likes **0**; sha `fa0bed22`). - 0★ HIGH still got a real card. Soft Noul ≠ hard safety. - `invented_signal: false`. - -- Hourly 0042 HIGH (`research/notes.md` §105): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#22**. Do **not** merge from - merged **#22** (hourly 2340 / §104). Do **not** re-fold 2340 - / §104 / 2246 / §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / - 1843 / §99 / 1740 / §98 / 1639 / §96 / - gliner-native-runtime / §97 / 1541 / §95 / jev-align - *mechanism* / §93. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating - displayed p as proof, schema-valid ordinary-model JSON - as a calibrated Noul, collapsing gulagala001/jevify into - Mintzs/jevify, quoting 40.3% without the constant-answer - baseline, hard-gating fastloop confidence as safety, - treating fan-out 0.0000 sd as universal determinism, - treating 2.95× RL as Harbor, hard-gating Laya cascade - 0.60, collapsing locate into decide, treating snake - points as intelligence, treating jevcheck as a - correctness proof, treating `no_direct_evidence` as - merge-safe, or treating a theme Choice as a music-theory - certificate is soundness theater. Hard-gating a soft - Noul as safety is the anti-pattern. Unique consecutive fragments: structured probability readouts; distribution > argmax; Noul 0.5 midpoint; score is expectation not integer; bare HTTP not SDK; Arohtea/jev-readout; Jev-style Choice/Score/Noul from ordinary models; optional DSH plugin; schema-valid ≠ calibrated; gulagala001/jevify ≠ Mintzs/jevify; Laya RLCD benchmark; 40.3% below constant-answer; open-weight measurement; mourad-ghafiri/laya-rlcd-benchmark ≠ yibie/laya-jev-lab; cheap fail-open semantic edge; second signal not sole; FastLoopError catch; SupremeDreamZ/jev-fastloop ≠ jev-ultrafast; asking more questions in one call; 0.980 at every N; nearly not fully deterministic; TheWebDevel/jev-fanout; Qwen3-VL perception + Jev decisions train RL; 0 model calls at deployment; VLM alone 1.7 vs +Jev 4.4; harneet2512/reflexrl ≠ khordoo/jev-reflex-autonomy-lab; independent Jev API vs Laya; cascade 0.60 matches 78% at 1.8×; noul facts not judgements; yibie/laya-jev-lab ≠ dairui1/jev-lab ≠ BrendanH18/jev-lab; GLiNER vs GLiFormer vs Laya vs Jev; extractors ≠ decision engines; Laya dict-instructions collapse 58.3%; umstek/zero-shot-ie-bench; decisions-per-minute & cost; 204 moves vs 73; throughput not intelligence; angelgalvisc/snake-arena-jev-vs-llms ≠ vtrivedy/jev-plays-games; behavioral contracts; pin expectations eval upgrades; raw 0.94 is not a release; sathariels/jevcheck ≠ dayhaysoos/jevals ≠ SivletLabs/jev-eval; evidence-linked dependency upgrade; Jev never generates filenames; no_direct_evidence ≠ safe to merge; GaneshVG18/upgrade-radar ≠ LYchoon/paper-radar-jev; discography theme/mood/complexity; five atomic questions one call; lirantal/discoprint. Twelve HIGH - clusters: **structured probability readouts** - ([Arohtea/jev-readout](https://github.com/Arohtea/jev-readout) - PRIMARY; structured probability readouts; distribution > - argmax; Noul 0.5 midpoint; score is expectation not - integer; bare HTTP not SDK; JavaScript; license null; - **0★**; HEAD `6f1e5900`; README SHA `67ee1e96`); - **ordinary-model Jev-shape** - ([gulagala001/jevify](https://github.com/gulagala001/jevify) - Jev-style Choice/Score/Noul from ordinary models; - optional DSH plugin; schema-valid ≠ calibrated; - gulagala001/jevify ≠ Mintzs/jevify; JavaScript MIT; - **0★**; HEAD `3d3e904a`; README SHA `0e2f8536`); - **open-weight Laya measurement** - ([mourad-ghafiri/laya-rlcd-benchmark](https://github.com/mourad-ghafiri/laya-rlcd-benchmark) - Laya RLCD benchmark; 40.3% below constant-answer; - open-weight measurement; - mourad-ghafiri/laya-rlcd-benchmark ≠ yibie/laya-jev-lab; - Python; license null; **0★**; HEAD `3401ff26`; README SHA - `d8d4859e`); - **cheap fail-open semantic edge** - ([SupremeDreamZ/jev-fastloop](https://github.com/SupremeDreamZ/jev-fastloop) - cheap fail-open semantic edge; second signal not sole; - FastLoopError catch; SupremeDreamZ/jev-fastloop ≠ - jev-ultrafast; Python MIT; **0★**; HEAD `1157841a`; - README SHA `053a0885`; GitHub size **12** (relock; was - **0** with contents)); - **fan-out measurement** - ([TheWebDevel/jev-fanout](https://github.com/TheWebDevel/jev-fanout) - asking more questions in one call; 0.980 at every N; - nearly not fully deterministic; Python MIT; **0★**; HEAD - `b30aaadc`; README SHA `394b2e1e`; GitHub size **206** - (relock; was **0** with contents)); - **VLM+Jev RL teacher** - ([harneet2512/reflexrl](https://github.com/harneet2512/reflexrl) - Qwen3-VL perception + Jev decisions train RL; 0 model - calls at deployment; VLM alone 1.7 vs +Jev 4.4; - harneet2512/reflexrl ≠ khordoo/jev-reflex-autonomy-lab; - Python MIT; **0★**; HEAD `aa36be84`; README SHA - `e6ff13cd`; default master; GitHub size **749** (relock; - was **0** with contents)); - **independent Jev API vs Laya** - ([yibie/laya-jev-lab](https://github.com/yibie/laya-jev-lab) - independent Jev API vs Laya; cascade 0.60 matches 78% at - 1.8×; noul facts not judgements; yibie/laya-jev-lab ≠ - dairui1/jev-lab ≠ BrendanH18/jev-lab; Python MIT; **0★**; - HEAD `30ba64dc`; README SHA `57bd1832`); - **locate vs decide** - ([umstek/zero-shot-ie-bench](https://github.com/umstek/zero-shot-ie-bench) - GLiNER vs GLiFormer vs Laya vs Jev; extractors ≠ - decision engines; Laya dict-instructions collapse 58.3%; - Python MIT; **0★**; HEAD `8770b16b`; README SHA - `d69dc96a`); - **throughput arena** - ([angelgalvisc/snake-arena-jev-vs-llms](https://github.com/angelgalvisc/snake-arena-jev-vs-llms) - decisions-per-minute & cost; 204 moves vs 73; throughput - not intelligence; angelgalvisc/snake-arena-jev-vs-llms ≠ - vtrivedy/jev-plays-games; Python MIT; **0★**; HEAD - `985a1c70`; README SHA `0db6f528`); - **behavioral contracts** - ([sathariels/jevcheck](https://github.com/sathariels/jevcheck) - behavioral contracts; pin expectations eval upgrades; - raw 0.94 is not a release; sathariels/jevcheck ≠ - dayhaysoos/jevals ≠ SivletLabs/jev-eval; Python MIT; - **0★**; HEAD `fc49c795`; README SHA `5c832f81`); - **evidence-linked upgrade review** - ([GaneshVG18/upgrade-radar](https://github.com/GaneshVG18/upgrade-radar) - evidence-linked dependency upgrade; Jev never generates - filenames; no_direct_evidence ≠ safe to merge; - GaneshVG18/upgrade-radar ≠ LYchoon/paper-radar-jev; - TypeScript MIT; **0★**; HEAD `e438f9bd`; README SHA - `c32f7d18`); - **knowledge-work discography** - ([lirantal/discoprint](https://github.com/lirantal/discoprint) - discography theme/mood/complexity; five atomic questions - one call; TypeScript Apache-2.0; **0★**; HEAD - `a9d3294f`; README SHA `9a64f473`). Soft Noul ≠ hard - safety. 0★ HIGH still got a real card (all twelve - **0★**). Census **live REST pulse** (independent review - relock after `0558f7d`): SemIf **2069★** (+22 vs §104 **2047**); jevlike - **1022★** (+4 vs §104 **1018**); TypeAR **12★** **flat**. - Tracker likes **59** (+3 vs §104 **56**); lastModified - `2026-09-20T04:29:16Z` **UNCHANGED** vs §104. Hub - `archerhume/4rcherhume` HTTP **401**. Archer still NOT - landed. Laya Hub likes **705**. Blackwood Hub likes **2** - gated manual — user census absent from tracker. Awesomejev - 561/27007 user-provided (≠ AnotiaWang/awesome-jev - **92★** ≠ yibie/awesome-jev **450★**). Qwen3.8-27B ≠ Archer (likes **15796**). - `invented_signal: false`. Do not copy keys / `npm` / - `pip` / `npx` / `uv` / `.env`. No wrapper. No invented - metrics. Relock: Arohtea HEAD `03734993`→`6f1e5900` / - README `76a99fe5`→`67ee1e96`; upgrade-radar HEAD - `1e91eb03`→`e438f9bd` / README `b7603c01`→`c32f7d18`; - discoprint HEAD `c2a3d47e`→`a9d3294f` / README - `392efd29`→`9a64f473`; yibie/laya-jev-lab size **49** - (was **0** with contents); umstek size **50** (was - **33**); jev-fastloop size **12** (was **0** with - contents); jev-fanout size **206** (was **0** with - contents); reflexrl size **749** (was **0** with - contents); upgrade-radar size **908** (was **0** with - contents); gulagala001/jevify size **145** (was **0** - with contents); mourad size **164** (was **0** with - contents); discoprint size **239** (was **0** with - contents); Arohtea size **44** (was **0** with - contents). Design claims unchanged. -- Hourly 2041 HIGH (`research/notes.md` §101): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#18**. Do **not** re-fold 1943 - / §100 / 1843 / §99 / 1740 / §98 / 1639 / §96 / - gliner-native-runtime / §97 / 1541 / §95 / jev-align - *mechanism* / §93. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating a - zero binary name gap as a fairness certificate, letting - Jev emit the verdict string, pasting PrismNLI's lead - without the contamination caveat, treating J7 pass as - safe to obey, treating a receipt as authorization, or - hard-gating confidence ≥0.95 is soundness theater. Nine - HIGH clusters: **resume-screening bias audit** - ([natemoo-re/bias-bench](https://github.com/natemoo-re/bias-bench) - PRIMARY; name×resume factorial independent Nouls; - callback determined by resume quality; mean-probability - name gaps operationally negligible; natemoo-re/bias-bench - ≠ BBQ; JavaScript; license null; **0★**; HEAD - `fe2f2535`; README SHA `a1c3e604`); **MCDA panel - code-owned verdict** - ([austindixson/planalyzer](https://github.com/austindixson/planalyzer) - Plan/PRD panel → code-owned pass|review|block; - cheerleading out of scope; Python MIT; **0★**; HEAD - `39fc161f`; README SHA `6e4d8da3`); **EU cost-aware - routing** - ([cannacre8ive/switchboard-ai](https://github.com/cannacre8ive/switchboard-ai) - cost-aware multi-model routing/escalation; decide vs do; - successful-task cost; JavaScript MIT; **1★**; HEAD - `5cae9d1c`; README SHA `872de837`; package **0.4.0**); - **frozen-protocol class bake-off** - ([elcronos/jev-vs-open-decision-models](https://github.com/elcronos/jev-vs-open-decision-models) - TypeSafe Jev vs PrismNLI vs Laya; contamination caveat; - Python; license null; **0★**; HEAD `b61e6cfc`; README - SHA `b7256888`); **VOI admission** - ([cvsgireesh/jevusher](https://github.com/cvsgireesh/jevusher) - context-window admission control; fail polarity per - lens; on small inputs lenses lose money; TypeScript MIT; - **0★**; HEAD `d830d344`; README SHA `428a4a59`); - **Leveson control plane** - ([MokiMeow/jev-fabric](https://github.com/MokiMeow/jev-fabric) - typed decision control plane; receipt ≠ authorization; - historical-v0 zero retained cases; Apache-2.0; **0★**; - HEAD `95b9a4f3`; README SHA `f485dbdd`); **scoring - economics** - ([jose-troche/live-rubric](https://github.com/jose-troche/live-rubric) - live 15-dim typed rubric re-score per pause; OpenJev/Codiv - ≠ TypeSafe hosted; ~$0.000004 desc / ~$0.000006 README; - TypeScript; license null; **0★**; HEAD `db8da8db`; - README SHA `4a0be084`); **pre-registered calibration - science** - ([willkelly/jev-evaluation](https://github.com/willkelly/jev-evaluation) - 28 predictions before data; 123,805 requests; confidence - does not track ignorance; polite injection 65% / crude - 0%; Python MIT; **0★**; HEAD `c168c093`; README SHA - `2d66ac22`; rh-guard owns injection); **class - infrastructure SDK** - ([nshkrdotcom/system_one_sdk](https://github.com/nshkrdotcom/system_one_sdk) - provider-neutral Elixir/BEAM Noul/Choice/Score SDK; - GitHub desc provider-neutral / README TypeSafe-first; - Elixir MIT; **0★**; HEAD `c2a522ee`; README SHA - `c117b4c4`; mix **0.5.0**). Soft Noul ≠ hard safety. - 0★ HIGH still got a real card (switchboard now **1★** - live REST). Census **live REST pulse**: SemIf **1984★**; - jevlike **1002★**; TypeAR **11★** flat. Tracker likes - **54** (+3 vs §100 pin **51**); lastModified **CHANGED** - 2026-09-20T02:59:13Z; Hub `archerhume/4rcherhume` HTTP - **401** (not re-fetched as a rewrite). Archer still NOT - landed. Awesomejev 561/27007 user-provided (≠ - AnotiaWang/awesome-jev 84★). - `invented_signal: false`. Do not copy keys / `npm` / - `pip` / `npx` / `uv` / `mix`. No wrapper. No invented - metrics. -- Hourly 1943 HIGH (`research/notes.md` §100): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#17**. Do **not** re-fold 1843 - / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime - / §97 / 1541 / §95 / jev-align *mechanism* / §93. - How-to-apply / mental models / architecture / - Harbor-jevals / toolbelt — not a thin Jev skill dump. - Backend-agnostic categorization/scoring/decision class. - Formal methods compose with scoring; a Noul is a SENSOR; - hard-gating `feels` 80%, pasting 17/18 as Harbor, - hard-gating 0 of 157, treating boolean @ 0.5 as a proof, - claiming 10×, treating softmax as a Noul, or treating 62 - tests as quality is soundness theater. Six HIGH clusters: - **Jev IS the if-statement** - ([southpolesteve/probably](https://github.com/southpolesteve/probably) - PRIMARY; judgments/probabilities drive branches; text - model only writes prose; interpreter owns - variables/loops/budgets/replay; otherwise maybe / - confidence gate; chaos samples after the gate; - southpolesteve/probably ≠ carldaws/hunch ≠ feelings ≠ - Kungie/gut ≠ Illusion47586/judge ≠ tidymodels/probably; - TypeScript MIT; **3★**; HEAD `6bf671a4`; README SHA - `c28570a9`); **GEPA live delta** - ([sutro-sh/jev-align](https://github.com/sutro-sh/jev-align) - 133★ / forks 10 live; build calibrated classifiers from - human feedback; HEAD `49753df9` / README SHA `363fccb7` - unchanged vs §93); **retrieve by relevance not - resemblance** - ([samdotmak/jev-recall](https://github.com/samdotmak/jev-recall) - one calibrated yes/no per memory in one request; pointer - mode 17/18 19/20 *theirs*; embedding resemblance misses - the allergy; MIT; **6★**; HEAD `d3e4acfb`; README SHA - `ea774519`); **memory leases ended by new evidence** - ([chopratejas/invalidate](https://github.com/chopratejas/invalidate) - HIGH upgrade; six Nouls then fixed rules in code; 0 of - 157 false invalidations; questions/plans/directives are - not evidence; unsure → review queue; host keeps the - store; Apache-2.0; **11★**; HEAD `d6ade601`; README SHA - `8cccad5f`); **contract-of-artifact lint rename** - ([mizchi/jev-lint](https://github.com/mizchi/jev-lint) - is mizchi/jevlint rename; name↔body / comment truth / - test-claims; no shipped rule has severity error; ~1 in 5 - findings wrong *theirs*; TypeScript MIT; **13★**; HEAD - `62d73f8e`; README SHA `4c0e37cd`); **JSON Schema - question compiler** - ([Kiln-AI/jev_jsonschema](https://github.com/Kiln-AI/jev_jsonschema) - HIGH upgrade; JSON Schema → typed JSON via Jev; - noul_threshold 0.5 decoder not a proof; - IncompatibleSchemaError lists every bad property; MIT; - **5★**; HEAD `fccea8c2` unchanged); **local System One - economics** - ([mizorewww/laya-coreml](https://github.com/mizorewww/laya-coreml) - on-device Laya CoreML ANE; ~5 ms P50 short decisions; - 189/189 FP16 checkpoint parity; 10× not achieved; - Apache-2.0; **0★**; HEAD `47f4baf0`; README SHA - `2068c661`) + - ([Micha0827/snapjudge](https://github.com/Micha0827/snapjudge) - softmax over allowed tokens ≠ Noul; question-first - cache; Python MIT; **3★**; HEAD `2df5ce27`; README SHA - `64a91458`) + - ([direwolfiy/JevPi](https://github.com/direwolfiy/JevPi) - Jev-first Pi agent loop; slow-LLM fallback; explicit - action menu / CandidateSource unimplemented; 62 tests - wiring not quality; license null; **0★**; HEAD - `980f8895`; README SHA `88ec1a55`). Soft Noul ≠ hard - safety. 0★ HIGH still got a real card. Census **live - REST pulse**: SemIf **1954★**; jevlike **989★**; TypeAR - **11★** flat. Tracker likes **51** flat; lastModified - UNCHANGED 2026-09-19T18:37:18Z; Hub - `archerhume/4rcherhume` HTTP **401** (not re-fetched as - a rewrite). Archer still NOT landed. Awesomejev - 561/27007 user-provided (≠ AnotiaWang/awesome-jev 83★). - `invented_signal: false`. Do not copy keys / `bun` / - `pip` / `npx` / `uv` / HF download. No wrapper. No - invented metrics. -- Hourly 1843 HIGH (`research/notes.md` §99): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#16**. Do **not** re-fold 1740 - / §98 / 1639 / §96 / gliner-native-runtime / §97 / - 1541 / §95. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; hard-gating - cost-derived 0.038 as a proof, merging von Needle - 52.6% with n=78 93%, “guaranteeing” calibration, or - pasting “Jev wins guardrailing” is soundness theater. - Five HIGH clusters: **cost-derived YES/NO/UNSURE - control flow** - ([Kungie/gut](https://github.com/Kungie/gut) - PRIMARY; cost-sensitive decision theory × System One - probabilities → control flow; thresholds derived from - costs not hard-coded; YES / NO / UNSURE from - cost_false_yes / cost_false_no / cost_human; - auto-batching same-object questions; Kungie/gut ≠ - tpellet/hunch ≠ carldaws/hunch; GitHub Apache-2.0 / - LICENSE MIT / pyproject Apache-2.0 *theirs*; **0★**; - HEAD `cb56c875`; README SHA `630474f6`; pre-alpha); - **typed-callback twin** - ([Illusion47586/judge](https://github.com/Illusion47586/judge) - judgment vs generation; deterministic execution after - probabilistic judgment; exactly one app-owned - callback; explicit uncertain branch; - Illusion47586/judge ≠ lexingtonhibiki/judgekit ≠ - Ascurse/typed-judge-kit; TypeScript MIT; **0★**; HEAD - `e69f65a1`; README SHA `08554c6f`; `@brkn-labs/judge` - 0.1.0); **variable-N option scoring as the trainable - object** - ([zwliJay/jev-forge](https://github.com/zwliJay/jev-forge) - dynamic candidate bags not fixed label sets; - zwliJay/jev-forge ≠ NanoJev; class-architecture note, - not a sixth species; GitHub NOASSERTION / LICENSE MIT; - **1★**; HEAD `eb3e4a2d`; README SHA `359f3f57`; do not - clone Hub weights); **open NAR replica economics** - ([wfzyx/von](https://github.com/wfzyx/von) late-catch - HIGH of the same repo; NAR local drop-in; open replica - economics / latency vs closed Jev; competing NAR - claims / replica honesty; Apache-2.0; **43★**; HEAD - `b9e42b26`; README SHA `574aa628`; do not merge Needle - 52.6% with n=78 93.0%); **typed vs chat judges on - guardrailing** - ([ishaannk/llm-vs-jev](https://github.com/ishaannk/llm-vs-jev) - cross-note only; nothing wins outright; can be argued - out of guarding; deeper integrity fold is rh-guard; - Apache-2.0; **0★**; HEAD `182e0864`; README SHA - `d9ebd40f`). gut/judge are **control-flow / - decision-theory overlays, not new class-table - species**. Soft Noul ≠ hard safety. 0★ HIGH still got - a real card. Census **live REST quoted**: SemIf - **1936★**; jevlike **983★** (watch claimed 984); - TypeAR **11★** flat. Tracker likes **51** flat; - lastModified UNCHANGED 2026-09-19T18:37:18Z; Hub - `archerhume/4rcherhume` HTTP **401**. Archer still NOT - landed. Awesomejev 561/27007 user-provided (≠ - AnotiaWang/awesome-jev 83★). X MCP `since_id` held; - pages_archived 0; no invented tweets. - `invented_signal: false`. Do not copy keys / `pip` / - `npm` / `uv` / Hub download. No wrapper. No invented - metrics. -- Effect-oriented loops (`notes.md` §28, `mappings.md` §19): Ward's - ZIO client keeps Jev as the outer Choice and the handler as the - effect. Not Effect.ts. GLiNER author: GLiNER2 "like jev" is GLiGuard - schema-conditioned categorize, not a Noul. -- Boundary-audit stop conditions for TOCTOU-of-Noul and vacuous specs; - FAQ rows for Alloy vs Apalache and PufferLib-as-DST-trio -- Research pointer to [dayhaysoos/jevals](https://github.com/dayhaysoos/jevals): - local MIT workbench for Jev questions vs labeled Noul/Choice/Score cases - (compare runs, WebMCP + agent skill). Empirical acceptance-test surface - for Hypothesis mapping cards; complements `evaluate_decisions.py`. Not a - jevals how-to (`research/notes.md` §24; one sentence in `validation.md`) -- Mental-models card: Augustus is design judgment across AI, SWE, - business, knowledge work, and life — not SWE-only. Pillars: expected - utility / selective classification, calibration and cost-sensitive - thresholds, VOI, MCDA, search/control substitutions, signal detection, - Leveson org/safety, NATM/snap-fit/Norman/Kent/Shirky as general - intuition. Domain gallery labeled Hypothesis except launch-week - Empirical SWE rows. -- Archer Hume architecture reconstruction (17 Sep 2026 essay, ~10k - probes of `jev-1.13.0`): direct readout vs generated confidence, - isolated questions, listwise IIA and order sensitivity, confidence as - arithmetic on the distribution. Independent envelope probe; does not - override live TypeSafe docs. Announced open-weight drop is **WATCH** - (27B dense, AU healthcare residency, prefers "decision models"; still - no Hub weights). `research/notes.md` §31, §33; `judgment-class.md` - when-to-use table; FAQ confidence / surfaces questions. -- Entropy as allocator (**Hypothesis**, `judgment-class.md`): Atallah's - low / medium / high buckets place System One on typed decisions and a - frontier decoder on high-entropy synthesis — same axis as marginals - vs joint and as VOI. "Review this PR" as medium is still partly - generative; "first model ever" is a claim. `research/notes.md` §38 -- Marginals, not a probabilistic program (`judgment-class.md`, FAQ): - Erik Meijer — Jev is a cool API and not a PPL; Kleisli qualifications - exaggerate; "Jev gives you the marginals; a decoder gives you the - joint." Joints and invariants stay with TLA+ / Alloy / contracts. - `research/notes.md` §34 -- Bespoke Nimble: open contrastive recipe, not a Jev distill. Model - card Apache-2.0 LoRA on Qwen3.5-9B (repo license absent). Their - 324-example holdout is a named receipt (Nimble 90.12%, Jev 1.13.0 - 93.21%), not a ranking. 9B-vs-Jev on your labels stays Hypothesis. - `research/notes.md` §35; one sentence in `validation.md` -- djev-spark: third compute graph (diffusion structured reads, - Jev-shaped I/O, image-in). Empirical as the public interface; - Hypothesis that it beats a decision head on your task. Archer's - multimodal drop stays WATCH. `research/notes.md` §36 -- Perception specialist then judgment specialist vs shared multimodal - System One (**Hypothesis**): SAM 3.1 (masks and tracks) or an ASR - transcript, then typed decisions on that state, is an application - pattern, not native omni. Information dies at the interface. Prefer - a shared multimodal decision model when the joint matters (Archer - Watch, not Empirical; djev-spark images; future audio). Basit ask, - primary post not retrieved. `research/notes.md` §39 -- Perception→decision pipeline, measure, and hill-climb - (**Hypothesis**, `validation.md`): stages with a versioned state - contract; stage metrics plus a frozen taskset; HoH changes one stage - or one interface. DSPy/Ax only on LM-program knobs; jevals and - calibration for the decision slice; Harbor names product - end-to-end, not a tutorial. `research/notes.md` §41 -- Eval & hill-climb (`validation.md`): jevals decision-stage hygiene - (independent keys, correctness is not confidence, held-out, immutable - runs) and Harbor as the product taskset substrate; one composition - table. `research/notes.md` §40 +- **Class breadth.** TypeSafe Jev remains the documented exemplar, not + the monopoly. Folded class peers and cousins: Laya (open head), kev + family (Archer-arch fidelity; replica honesty), OpenJev `/v1/decide` + (not a TypeSafe drop-in), TypeAR / constrained-AR, GLiNER species + (locate vs categorize vs safety-schema vs local multi-head), Decision + Graph Protocol (frame→assess→commit; app retains permissions/effects), + encoder zero-shot classifiers. +- **Encoder / ZS lineage (Merve Noyan, `notes.md` §112).** Institutional + HF voice: BERTForXYZ → DeBERTa → ModernBERT. Many problems solved with + LLMs could have been solved with zero-shot classifiers — it was a + skill issue. Prefer DeBERTa and ModernBERT heads. Jev vs GPT-5.6 + bakeoffs are a category error. Softmax / ZS scores still ≠ calibrated + Noul; soft scores ≠ hard gates. Multimodal image↔text ZS is a + perception front-end, not a decision model. Quote *theirs*; no + invented accuracy numbers. +- **Measurement honesty.** Harbor / jevals practice: ranking ≠ calibration; + Score is a 0..n−1 expectation, not 0–1; Noul has no confidence field; + ECE ≠ an edge; treating 0.85 / minProbability as a hard Harbor gate is + theater; soft Noul ≠ hard gate; VERIFY must acquire discriminating + evidence, never same-pool confidence-only rescoring. Calibration is + not alpha. Compaction default 0.5 is not safety. +- **Composition / placement.** Decision Graph Protocol envelope; + meaning-grep (proposition ≠ embedding; AND/OR/NOT after threshold; + not a gate); gut cost-of-error overlay (thresholds from costs, not + hard-coded); fail-open vs fail-closed per action; prune ≠ deny; + hard-gating DGP as safety theater. +- **Pages landing** at https://24601.github.io/Augustus/ (`docs/`) plus + community-health stubs (`SECURITY.md`, `CONTRIBUTING.md`, brief + `CODE_OF_CONDUCT.md`), `CITATION.cff`, and + `.github/workflows/scorecard.yml` so an OpenSSF Scorecard can appear. + No invented Scorecard number. +- Formal methods pillar unchanged in spirit: a Noul is a SENSOR; never + launder it as a proof (soundness theater). ### Changed -- Skill description rewritten as trigger conditions (mixed architecture, - prefilter, routing, preference lint, classification skepticism, family - choice including GLiNER/GLiClass/listwise/vision) plus an explicit `not_for` - against the official `typesafe-ai` skill -- Identity lock vs neighbor skills (`typesafe-ai`, `tenbin`, `decision-first`) - so Augustus stays the design-judgment layer — class-wide, not TypeSafe-only -- Design cards name hole, family, and typed judgment provider (Jev default; - other family only with self-eval) -- Protocol fan-out step is family-aware (Jev batch, GLiClass one-pass, - dual-encoder prompt scoring); ranking vs decision fail policy is a - non-negotiable -- Protocol and FAQ branch for "formally verify with Jev"; methods-catalog - and composition-algebra verifier position point at the ownership split -- Skill mission and description are domain-general (AI / SWE / business / - knowledge work / life); FAQ "is this only for software?"; mappings.md - beyond-SWE examples labeled Hypothesis; boundary-audit red flags for - TOCTOU-of-Noul and vacuous specs; formal-methods expanded with Alloy vs - Apalache and the DST trio including PufferLib; GLiNER promoted from - cousin footnote to species-map peer - -### Fixed +- Marketplace plugin version and SKILL YAML pin: 0.3.0 → 0.4.0. +- README / Pages SEO: one-liner, homepage, install paths, "not a + TypeSafe product" clarity, [`rh-guard`](https://github.com/24601/rh-guard) + companion pointer. +- CHANGELOG restored to Keep a Changelog. The hourly uniqueness dump + lives in `research/changelog-hourly.md`. -Adversarial review of the whole skill against its own non-negotiables -(findings in `research/notes.md` §27). - -- Gate fail policy is per action, not universally open - (`composition-algebra.md` position 3, `agent-self-assessment.md`): - advisory guards fail open *because* an interlock sits underneath; - selection and authorization gates fail closed -- Dual-orchestration topology A selects from a closed catalog instead of - "planning" MCP calls, which contradicted the standing planner rejection -- Species map applied to the skill's own advice: GLiClass (categorize) is - the large-catalog substitute for a 255-option Choice; GLiNER spans are - not (`SKILL.md`, `judgment-class.md`, `applied-mappings.md`) -- Han Xiao trolley relabeled an Empirical **rejection** (one tweet, no - repo), not a recipe -- openjev-lm caveat moved to the figure it belongs to: 92.9% is against 70 - hand-labelled gold, 98.1% is teacher *agreement* -- Contract surface removed from design cards: the Ax constructor call and - the `instructions` key enumeration point at live docs instead - (`optimizer-integration.md`, `question-design.md`) -- `mappings.md` preamble no longer claims uniform Hypothesis where card - bodies say Contract/Empirical; §17 forbids reusing jevgate's ≤0.18 as a - constant; all Hypothesis-range references aligned to §6–§19 -- Ownership split labeled Contract in `toolbox-mapping.md`, matching - `mappings.md` §8; done-check splits structure from the Noul +### Security -Second pass on `7b3a0c3` (`research/notes.md` §43). Zero blockers. -Dropped the unpublished `npx jevals` line; SAM and ASR are upstream -producers, not the perceive species; removed two call shapes from -`optimizer-integration.md`; tagged the $0.042/MTok cell as a vendor -figure; marked GodsBoy 94.4% exploratory. -- Skill description gained trigger terms for boundary audit, question - diagnosis, agent self-supervision, and optimizer placement +- `SECURITY.md`: supported line is 0.4.x; report via GitHub Security + Advisories / private vulnerability reporting. Dependabot security + updates, secret scanning, and push protection stay enabled. +- `.gitignore` covers common secret filenames. -## [0.2.0] - 2026-09-18 +## [0.3.0] - 2026-09-18 Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) (`65a39f3`, 2026-09-12). ### Added -- Boundary-audit card for existing systems: three-way split (exact / - bounded judgment / generation), code-smell catalog, fit test, opportunity - map, smallest-viable-boundary rule, Jev-around-LLM sandwich, centralized - policy + raw-judgment retention, red flags, completion questions -- Protocol branch: audit a codebase/PR before inventing mappings; per-action - risk gates; keep questions/thresholds in one reviewable module -- Skill description trigger terms for brittle parsers, prompt-to-JSON - classifiers, and agent loops that are really bounded decisions - -## [0.1.0] - 2026-09-18 - -Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) -(`65a39f3`, 2026-09-12) — the only tagged revision of the official skill at -Augustus launch. - -### Added - -- Skill protocol, decision-design card, and evidence labels (Contract / - Empirical recipe / Hypothesis) -- Classical-method mappings: features/utility, selective decisions, decision - circuits, bounded rerank, hierarchy/beam search -- Agent self-assessment, optimizer coupling (Ax, DSPy, ProgramAsWeights) -- Toolbox sweep, named-methods catalog, 11-position composition algebra, - question-design diagnosis -- Validation gates and `scripts/evaluate_decisions.py` -- Launch-week evidence archive (187 repos) and public ecosystem index -- Claude Code marketplace manifest; public GitHub mirror at - [`24601/Augustus`](https://github.com/24601/Augustus) - -### Research log (pre-tag) - -The dated passes below are how 0.1.0 was assembled. - -### 2026-09-18 (refresh pass 2) -- Research: Gemini Deep Research report retrieved and archived (interaction ID - saved in jev-archive/state); GitHub census doubled to ~60 Jev repos + - framework integrations (LiteLLM, LangChain, Vercel AI, Mastra, Eliza, Ax, - Composio); all 87 repos cloned to /home/user/workspace/jev-archive for - hourly refresh. -- New measured recipes added to notes.md: foreman supervision loop, pi-jev - gate thresholds, pi-warden 6→0 paired-run result, winnow relevance sieve, - fast-jev-compaction two-noul rule, skill-router gates (0.30/0.40, shortlist - 3, 94.4% vs 70.8%), calibration ECE 0.0313 vs 32% OOD collapse (Archer - Hume), Every 777-judgment eval, Near Here moderation numbers. -- Skill: added references/agent-self-assessment.md (agent self-supervision - lifecycle, grounding/citation checks, skill callability testing) and two - mapping-index rows; validation.md dogfooding section still canonical. - -### 2026-09-17/18 (initial) -- Baseline research archive (sources.json, notes.md), augustus skill with - mappings + validation references, evaluator script, hourly refresh script, - Claude plugin marketplace manifest. - -### 2026-09-18 (topic-index pass 3) -- Fixed census method: exact GitHub search paginated (700 repos created since - 09-14 captured; 700-result cap noted) + topics/jev crawl → ~80 additional - repos; archive now 184 clones. Miss-cause documented: earlier star-sorted - limit-40 search cut the low-star tail (incl. both MCTS repos). -- Skill: MCTS mapping promoted experimental → empirical recipe (grounded vs - speculative fidelity in types; probes-only concession; measured 24/24 vs - 1/24 greedy); agent-self-assessment.md gains the judge-variance recipe - (Jev judge 224-279x more consistent than LLM judge over 100 reps). - -### 2026-09-18 (pass 4 — optimizers + official skills + clone audit) -- ax Jev support documented from source (native adapter details, trueThreshold - semantics, fail-closed mapping validation); new reference - optimizer-integration.md covering Ax + DSPy typesafeify + jev-dspy-lab. -- typesafeainate/dspy-typesafeify cloned; official typesafe-ai/skills already - archived and layered-on (never duplicated). -- Clone audit: repos.txt deduped (185 unique), 0 missing on disk, no failures. - -### 2026-09-18 (pass 5 — toolbox sweep meta-method) -- New references/toolbox-mapping.md: the how-to-find-approaches-and- - applications procedure (judgment-shaped-hole substitution, newly-feasible - classification via economics inversion, standing rejections list); wired - into SKILL.md central model + index row. - -### 2026-09-18 (pass 6 — named-methods + operators/theorems tier) -- references/methods-catalog.md: ~20 named algorithms (CatBoost row is - Empirical via autoresearch cookbook) + operators/theorems tier with - precondition-carrying rule; wired into SKILL.md index and toolbox sweep. - -### 2026-09-18 (pass 7 — composition algebra as application generator) -- references/composition-algebra.md: 11-position grammar of Jev-vs-construct - relations, logical-operator combination rules, and the position×construct - traversal as the systematic application generator; wired into SKILL.md - index + toolbox sweep. -- Hourly 2340 HIGH (`research/notes.md` §104): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#21**. Do **not** re-fold 2246 - / §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 - / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime - / §97 / 1541 / §95 / jev-align *mechanism* / §93 / - jev-orderby-bench *six-gates* / §60. How-to-apply / - mental models / architecture / Harbor-jevals / toolbelt - — not a thin Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating a - description-only stub as a checkpoint, re-folding - six-gates as new, distilling Jev as teacher of record, - pasting openjev-sglang as OpenJevPro, treating - constrained logprob as a Noul, quoting an - untrained-looking demo as Jev identity, pasting a - radar's listed numbers, or quoting 0.916 as a class - ceiling is soundness theater. Distill-Jev UI stub and - constrained-logprob-as-Noul are the anti-patterns. - Eleven HIGH clusters: **from-scratch calibrated - decision model** - ([hyusi2003/MiniSystemOne](https://github.com/hyusi2003/MiniSystemOne) - PRIMARY; train calibrated ~27M from scratch; typed Q→prob dist / one forward pass / no LLM decode; - hyusi2003/MiniSystemOne ≠ Colvin0315/MiniSystemOne; description-only stub / size 5; - Apache-2.0; **0★**; HEAD `4385335b`; README SHA `83016bf8`); - **ORDER BY ranking upgrade** - ([yodablocks/jev-orderby-bench](https://github.com/yodablocks/jev-orderby-bench) - ESCI hard probe fails four of six; jev_bool ECE 0.242 inversion 0.255; - do not re-fold §60 six-gates as new; jobbyjev one-request-per-company from batch-size result; - Python MIT; **0★**; HEAD `52397954`; - README SHA `7bd075c3`; size **309**); - **find/design/evaluate decision loops** - ([karanb192/jev-architect](https://github.com/karanb192/jev-architect) - find/design/evaluate TypeSafe Jev decision loops; karanb192/jev-architect ≠ samtay32/jev-system-architect; - HTML MIT; **0★**; HEAD `35ea6d93`; README SHA `68c2b5f9`; size **5199**); - **distill-Jev UI stub** - ([Jairik/jev-distiller](https://github.com/Jairik/jev-distiller) - Jairik/jev-distiller size 1; distill-Jev UI stub / do not distill Jev as teacher of record; - MIT; **0★**; HEAD `0589d44c`; README SHA `aa408c5e`); - **post-launch scored opportunity map** - ([licensedsaucer9-web/jev-opportunities](https://github.com/licensedsaucer9-web/jev-opportunities) - post-launch scored use-case map / Jev self-scores then human curation; - license null; **0★**; HEAD `a47fa414`; README SHA `aa33f901`; size **27**); - **Jev-inize a use case** - ([gavinHuang/jevinize](https://github.com/gavinHuang/jevinize) - Jev-inize a use case into classifier/router; gavinHuang/jevinize → simple-jev not TypeSafe; - featherless-ai/simple-jev; - MIT; **0★**; HEAD `6d080632`; README SHA `5f48e622`; size **6**); - **saved-decision regression** - ([VihaanAgarwal/jev-diff](https://github.com/VihaanAgarwal/jev-diff) - compare saved decisions / same label can still change the branch; - VihaanAgarwal/jev-diff ≠ Saik0s/diffusiongemma-jev-macos; - not tested with a live Jev API key; Python MIT; **0★**; HEAD `a3c98807`; - README SHA `3b0ce75c`; size **120**); - **constrained-logprob API** - ([zhangcy122/OpenJevPro](https://github.com/zhangcy122/OpenJevPro) - constrained logprob + temp/Platt ≠ Noul; OpenJevPro pastes openjev-sglang JevBench as own; - zhangcy122/OpenJevPro ≠ IamBusy/OpenJev ≠ ekzhang/openjev-sglang; - PolyForm Noncommercial; HTML; SPDX NOASSERTION; **0★**; HEAD `94d77bcb`; - README SHA `50c77ace`; size **64** (relock; was **62**)); - **SmolLM RLCD reproduction** - ([patelvishwa112/jev-system-one-rlcd](https://github.com/patelvishwa112/jev-system-one-rlcd) - SmolLM-135M / sub-70ms / 0 output tokens; demo P(True) 0.5052 / Choice conf 0.2872 / Score conf 0.0055; - README claims MIT / GitHub license null / no LICENSE file; - patelvishwa112/jev-system-one-rlcd ≠ arnabgho/rlcd-lite ≠ blackwood-rlcd; - Python; **0★**; - HEAD `62b103b3`; README SHA `55994d69`); - **source-backed Awesome radar** - ([logicrw/awesome-jev-projects](https://github.com/logicrw/awesome-jev-projects) - source-backed Awesome Jev radar / 306+ commit-pinned; auto GitHub sync / Issue-only submissions; - logicrw/awesome-jev-projects ≠ AnotiaWang/awesome-jev ≠ yibie/awesome-jev ≠ cobanov/awesome-jev ≠ rupeshpoojary9/awesome-open-system-one; - JavaScript MIT; **136★**; HEAD `97057cc1`; README SHA `25a19b31`; size **7677** (relock; was **7136**)); - **rival-aware one-pass scorer** - ([olanotolu/jevbetter](https://github.com/olanotolu/jevbetter) - hashed n-gram encoder / rival-aware attention; olanotolu/jevbetter vs jevlike starter; - synthetic hard menus top-1 0.916 vs 0.873 / ECE 0.0182 vs 0.0367 / 40 vs 4608 menus/sec; - shuffled-context control 0.335; Python MIT; **12★**; HEAD `bb0ebc82`; - README SHA `5cbe01d4`; size **324**). - Soft Noul ≠ hard safety. 0★ HIGH still got a real card - (awesome-jev-projects **136★**; jevbetter **12★**). Census - **live REST pulse**: SemIf **2047★** (+28 vs §103 **2019**); - jevlike **1018★** (+12 vs §103 **1006**); TypeAR **12★** - **flat**. Tracker `multimodalart/jev-reproductions-tracker` - likes **56**; lastModified - `2026-09-20T04:29:16.000Z`; Hub `archerhume/4rcherhume` - HTTP **401** (not re-fetched as a rewrite). Archer - still NOT landed. Awesomejev 561/27007 user-provided - (≠ AnotiaWang/awesome-jev **91★** ≠ logicrw **136★**). - Qwen3.8-27B ≠ Archer (likes **15787**). - Independent adversarial review relock (PR #22 - after `babb111`): HIGH HEAD/README unchanged; - star/size lag locked (awesome **136★**; SemIf - **2047★**; jevlike **1018★**; AnotiaWang **91★**; - OpenJevPro size **64**; simple-jev **318★**; - openjev-sglang **205★**; RLCD size **1513**; - Colvin size **814**; yibie **430★**; cobanov **224★**). - `invented_signal: false`. Do not - copy keys / `npm` / `pip` / `npx` / `uv` / - `.env`. No wrapper. No invented metrics. - Mixed-architecture card: default placement is judgment-class model + generator + code, not stack replacement. Covers cost-sensitive prefilter (fail-open vs fail-closed per action), tool/skill routing, AGENTS.md @@ -1457,1876 +166,6 @@ The dated passes below are how 0.1.0 was assembled. two-layer finish gate. pi-jev (not pi-jev-context). jev-plays-games option-order probe. joxide jump-by-description. laya-typed-decisions companion packaging. No wrapper. -- Hourly ~12:58 Boise fold (`research/notes.md` §44): Archer still - Watch (no architecture rewrite). Store-index fork: in-engine - ([sqlite-jev](https://github.com/mgaitan/sqlite-jev), pg-jev cousin) - vs CLI rewrite (jevql). Soft judgment inside a hard envelope - ([bitrate-advisor](https://github.com/affirmitv/bitrate-advisor); - mmalisper JOB planner +12% geomean, author-reported; join-order - Choice alone was 2× slower). Distill-to-device as a *memory* gate - (jev-gate, already §33). Encoder vs decoder open-replica receipts - (openjev-lm $0/call overnight CPU). jev-harness as Harbor-adjacent - practice (assert on action). Host adapter - ([jev-routing](https://github.com/nekowasabi/jev-routing), not MCP); - OpenClaw typed routing ([jev-claw](https://github.com/trietphan/jev-claw)). - Voice-control and JevML are README stubs. Higgsfield auto-routing is - a claim. No wrapper. -- kev (`jaredpalmer/kev`, `research/notes.md` §45): runnable Archer - reconstruction on the trained decision-only open path next to Laya / - Nimble / Watch. Qwen2.5-0.5B LoRA + pointer, Apache-2.0, `POST - /v1/systemone` drop-in. Isolation exact (packed vs separate max Δ - 3.7e-6; secret-in-sibling p=0.03 vs in-state 0.99). Held-out ECE - 0.065 (0.031 after temp scale); acc 0.799 on 1,350 ID questions. - Permute argmax flips 7.4%; IIA log-odds shift mean 0.13; boundary - forgery held. Laptop-local System One for development/eval; not a - knowledge/frontier substitute; not a Jev teacher-copy. Contrast vs - TypeAR, encoder DeBERTa, proprietary Jev. jevals/Harbor bake-off - candidate. No serve how-to. -- Hourly ~14:03 Boise fold (`research/notes.md` §46): Archer still - Watch. Open multimodal RLCD - ([blackwood-rlcd](https://huggingface.co/BlackwoodAI/blackwood-rlcd), - CC BY-NC): screenshot + marked candidates → Choice; web acc 0.907 vs - Jev 1.13 text-only 0.480; letter-shuffle 0.133 vs 0.587; ECE 0.037; - ~200 ms H100; Jev still leads general text 0.850 vs 0.786. Shared - bake-off ([open-jev-laya-bench](https://huggingface.co/datasets/pngwn/open-jev-laya-bench)): - 26+9 tasks, 11959 items; ECE/NLL/Brier; macro acc Δ +0.023 - neutral / +0.229 home; LLM-as-judge is not the score. Decision-token - QLoRA - ([Foodoo1/Qwen3-14B-RLCD-Decision-LoRA](https://huggingface.co/Foodoo1/Qwen3-14B-RLCD-Decision-LoRA)): - fraud_risk 64→95%, overall 85.2→98.8% at ~234 ms/4-field broadcast; - synthetic. jevgate frame: allowlist *proves*, Jev judges only - unlisted, fail-open. wellposed: missing `other` → confidence 1.00 - wrong; gating cannot catch it (`tenbin` owns the lint skill). - S1 reflex keeps control (jev-reflex-autonomy-lab). MED: - jev-decision-layer, jev-e2e, jevpandas. No wrapper. -- Abide (`coldteadotai/abide`, `research/notes.md` §47): productized - Jev preference lint for Claude Code / Codex / OpenCode. Soft - AGENTS.md / CLAUDE.md rules → one Score per rule on the diff (never - the conversation); hard rules stay with the linter (same layering - family as jevgate). Edit- vs turn-phase observation window; banded - confidence (≥0.8 repair / 0.5–0.8 note / <0.5 silence — their - operating point) + fail-open hooks; rubric.json quotes source - lines; calibrate/tune fix false positives in the question. Replay - of 93 sessions (1,256 edits / 147 turns) with independent review: - edit precision ~26%, turn ~73% (author-reported, before tune). - Fuller productized path of the jev-pref contract. Complementary to - rh-guard (eval-integrity vs project soft rules). Text/diff only — - not multimodal. No hook how-to. -- kev delta (`research/notes.md` §45): Hub weights - [`jaredpalmer/kev-0.5b`](https://huggingface.co/jaredpalmer/kev-0.5b); - `--run` accepts Hub ids; PEFT `task_type=FEATURE_EXTRACTION` (publish - patches legacy adapters). HIGH question-design: confront Choice - `"other"` / none-of-the-above as a wrong alternative too, vary - wording, dedicated `none_of_the_above` eval (no published rates). - Cross-link wellposed request-shape lint. No species change. No - wrapper. -- Hourly ~14:52 Boise fold (`research/notes.md` §48): Archer still - Watch. Extractive selection + offline `redecide` - ([testimonial-miner](https://github.com/AppitStudio/testimonial-miner)); - pointer-not-generator - ([jev-reviewer](https://github.com/choxos/jev-reviewer)). Local - `/v1/systemone` drop-in ([jev-local](https://github.com/us/jev-local); - default scorer is a stub until `hf`). Observe→decide→verified-act, - no screenshots ([solari-reflex](https://github.com/hitakshiA/solari-reflex); - 60.2/194.9, 66/460, 24.2/98.4 s vs Codex on Solari). Dataframe - accessor sibling ([jevframe](https://github.com/ktaletsk/jevframe); - note jevpandas). Route ≠ memory (jev-hermes). Advisory sidecar - (agent-workflow-typesafe-ai). Structure induction (dag-jev experiment). - Decision-for-control / generator-for-content (jev-agentworld-web-simulator). - Collab arms + Wilson/McNemar (jev-testbench). AST ∩ semantic (jevscan; - `tenbin` owns lint). Light Pi gate (pi-jev-approver). Laya ONNX port - ([laya-onnx](https://huggingface.co/Mattepiu/laya-onnx); do not copy - vs-Jev table). Spotcheck: SemIf 1551★; jevlike 866★; tracker - 20:12:57Z still lists Laya, not Blackwood. No wrapper. -- Hourly ~15:52 Boise fold (`research/notes.md` §49): Archer still - Watch. X discourse blocked. Boundary map / extractable-from-state - ([jev-capability-atlas](https://github.com/Zaious/jev-capability-atlas); - history suite A wrong@0.90 / B 0.07 / C right@0.97; component node; - dangerous-high ECE; DOM-as-text + fan-out). Harbor-style bake-off vs - constrained LLMs - ([DMB](https://github.com/nibzard/decision-model-benchmark) v2: jev - banking 76.3% / spam 93.0% / 256+ cap; p50 264–276 ms; $0.07/1k; no - class wins on quality). Feedstock - ([jevals-data](https://github.com/Jevals/jevals-data) CC-BY-4.0; - recompute-from-logs; 2026-09-18 board). Dual-process S1 decide / S2 - generate ([dual-process-ai](https://github.com/taro1985/dual-process-ai); - routing accuracy unmeasured). Combinatorial ≠ extractive (ARC-AGI - Direct Jev 4/400). Packed one-forward open LLM - ([open-alternative-jev](https://github.com/ikermoel/open-alternative-jev) - RACE-H 92.9% @ 4.55 q/s; not a Jev reproduction). Tiny SAN local - surface ([von](https://github.com/wfzyx/von) 14 MB; not a replica). - kev light delta **100★**. Do not merge Banking77 87% / 76.3% / - 79.67%. No wrapper. -- GLiNER2.5 extractive compaction (`research/notes.md` §50, - [gliner25-compaction](https://github.com/m-newhauser/gliner25-compaction), - Apache-2.0): architecture notes, not a plugin how-to. Pointer - keep-drop (character-offset copies) vs generator summarizers; - family with testimonial-miner / jev-reviewer. Soft retention Choice - under a hard mutation envelope (mutating tools / shell operators → - `keep_full`); low-confidence / invalid evidence fail closed to - `keep_full` — contrast many fail-open Jev gates. Same compaction - *job* as fast-jev-compaction / pi-jev-compaction; GLiNER encoder - backend; Fastino/GLiGuard sibling class. `shadowMode` default true. - Not Jev. Not multimodal. No invented metrics. -- CI merge-gate / fail-open wake VOI / S1 indexer / claim-evidence - (`research/notes.md` §51): architecture notes, not a plugin how-to. - [latch](https://github.com/CaseReed/latch) cluster-then-policy - PASS/BLOCK (pair Harbor + rh-guard). - [wakegate](https://github.com/shitianfang/wakegate) skip only if - p(wake)<0.2 (21/21 smoke). s1-graphify-indexer GLiNER extract + - escalate-S2 (10–50× unfilled). - [clear-head](https://github.com/VladyslavHontar/clear-head) - claims vs session evidence. - reification-labs/foreman description-only Phoenix scaffold (not the - super-jev loop). - [jev-gateway-bench](https://github.com/vinilana/jev-gateway-bench) - Harbor on/off one-run signal. - jev-marshal Watch/empty; jevons bounded Pi supervisor (shadow - recovery). MED: if-ai, omp-auto-mode, downloads-sorter, label-desk, - herdr-jev. Archer still Watch. No invented metrics. No wrapper. -- GLiNER2 Ultrafast observe→score→act (`research/notes.md` §52, - [gliner2-ultrafast](https://github.com/sahibzada-allahyar/gliner2-ultrafast), - MIT): architecture notes, not a browser-agent how-to. Same - observe→score-among-candidates→code-acts *job* as jev-ultrafast / - solari-reflex; local GLiNER2 (`fastino/gliner2-multi-v1`) backend, - not GLiNER2.5. No screenshots; no generated selectors; code owns - actuators. Hybrid local decide + remote fill (Mercury 2.5 default - for TYPE). `DONE` ≠ verified success. Contrast blackwood-rlcd - screenshot multimodal; laya-mind2web is DOM-index Laya (same - observed-candidate family). Fastino sibling class with - gliner25-compaction (different hole) and GLiGuard (safety schema). - Demo (theirs, not re-run): Flights 12.20 s / 13.785 s / ~$0.0001 - API — demonstration, not a bake-off. No invented metrics. -- jev-pruner evidence-preserving Bash stdout prune (`research/notes.md` - §53, [jev-pruner](https://github.com/tamaratran/jev-pruner), MIT): - architecture notes, not a plugin how-to. After Bash, Jev Noul-prunes - stdout chunks before the main LLM sees them — no summary. Hard - envelope (≤10k estimated tokens / JSON-diff-whole-doc untouched) - then soft Noul; fail-safe keep original; full archive. Marketplace - id still `fast-jev-output`. Codex is opt-in wrapper, not automatic - interception. Same evidence-preserving *family* as - fast-jev-compaction and gliner25-compaction; different *job* - (command output vs session memory) and Jev backend vs GLiNER2.5. - Manual sweep (theirs): needles 24/24; mean reduction 83% on trim - scenarios. Harbor plugin-eval cannot reach Jev. Terminal-Bench - paired pilot is integration, not a full bench. No invented metrics. -- Cua-S1 specialist System One computer-use (`research/notes.md` §54, - [cua-s1](https://github.com/trycua/cua/tree/main/libs/cua-s1), - parent MIT, ~23.3k★ this pass): architecture notes, not a Driver / - MCP / `uv` how-to. Form-oriented profile `cua-s1-form-v0`. Byte - encoder + option-attention head chooses fill/check/click/skip per - observed element; does not generate values or selectors. Plan ≠ - execute; dry-run default; `execute`/`submit` independent opt-ins; - fail-closed on unknown checkbox / fill without advertised token - `set_value`. **Not TypeSafe Jev** — parallel "System One" naming in - CUA research. Same observe→score-among-candidates→code-acts *job* - as jev-ultrafast / gliner2-ultrafast / solari-reflex / laya-mind2web; - specialist form contract, source-only this pass (no weights, no - checkpoint scores). Offline metric *names* only (accuracy, - abstention, coverage, wrong actions/targets, unsafe when should - abstain). Tests exercise implementation, not checkpoint quality. - Watch for a `cua-s1-form-v0` artifact drop. No invented metrics. -- Hourly ~17:48 Boise fold (`research/notes.md` §55): Archer still - Watch. X MCP flap; `since_id` not advanced. Architecture notes, not - a how-to. Local CUDA/PyTorch Choice/Score/Noul replica - ([jevify](https://github.com/Mintzs/jevify); uncalibrated - likelihoods ≠ Noul; no LICENSE this pass; independent of - Distillation). Decision-native RAG - ([decision-native-rag-skills](https://github.com/emergency-lee/decision-native-rag-skills); - retrieve wide → decide → evidence set; no bundled harness; no - universal benchmark). Verbatim session ledger + scored recall - ([carryforward](https://github.com/Dharundp6/jev-carryforward); - rules never judged; fail-open dump; 9×3 hint). Judgment as a - Ruby language primitive ([hunch](https://github.com/carldaws/hunch); - English-as-config; `rescue nil` fail-open at save). Healthcare - Harbor-shaped S1+S2 - ([explore-typesafe-ai](https://github.com/si618/explore-typesafe-ai); - synthetic FHIR; not clinically validated). Pre-registered - independent eval - ([jev-baselines-eval](https://github.com/ickma2311/jev-baselines-eval); - **both AMBIGUOUS**; cascade sign-flip at exact parity; - confidence=1.0 theater; encoder-with-labels wins; serving-path ≠ - model-speed; same-day errata ×3). Student-b light delta only (HF - card unchanged). MED: toolgate (pre-exec allow/block/review; Jev - not authorization), typesafe-screening-mcp (PubMed screening aid), - databricks-jev-pdf-lab (**honest negative**; no OSS license), - yannip1234/codex-jev (extractive compression family; equal - accuracy/lower cost not established), kazuhideoki/jev-search - (recursive *file* search + fzf; **not** superagents-lab web - search). No wrapper. No invented metrics. -- Hourly ~18:38 Boise 2026-09-18 / 00:38 UTC 2026-09-19 fold - (`research/notes.md` §56): Archer still Watch. Architecture - notes, not a plugin / showcase catalog. Classify-first MCP - ([jev-sift](https://github.com/kbhuw/jev-sift); batch path/url/text - → Jev without entering main agent context first; 50 / 60k / 2MB / - public-IP envelope; mocks ≠ accuracy; no LICENSE this pass; - topology A MCP, not jev-routing). Living applied-mappings atlas - ([jevable.com](https://jevable.com/); claimed 342 vs JSON-LD first - page 36; class patterns — intent columns, score-among-observed, - VOI gates, generative UI decide, robotics text-state, draft-gate - silence ≠ safer — not a 342-title hit list). Maker clocks stay - claims unless already a named receipt. No wrapper. No invented - metrics. -- Stagehand experimental Jev stack (`research/notes.md` §57, - [#2955](https://github.com/browserbase/stagehand/pull/2955) 5/5 of - #2951–#2955, all OPEN draft): architecture notes, not an SDK - how-to. Major harness productization of - observe→score-among-candidates→code-acts (cousins jev-ultrafast / - gliner2-ultrafast / cua-s1 / solari). Jev picks a11y elements; - code copies text. extract `"off"` | `"judge"` | `"pick"`. Their - card (gemini-3.8-flash, 25×3): **37/75** no-LLM ~0.5 s vs baseline - **4.37 s**; 69/75 vs 23/25 (92% both); LLM-off **36/75** — pick is - a fast path, not a replacement. Screenshot extract always LLM. - Cache-check errors never block replay. Do not merge clocks. No - invented metrics. -- Hourly ~18:46 Boise 2026-09-18 / 00:46 UTC 2026-09-19 fold - (`research/notes.md` §58): Archer still Watch. Architecture - notes, not a Convex / uv / pnpm catalog. Public judgment wall - ([ask-jev-ai](https://github.com/waynesutton/ask-jev-ai); 6 - parallel questions; policy-in-code; cost-to-1M from tokens; - license null). Meaning-search without embeddings - ([jevgrep](https://github.com/Bentlybro/jevgrep); 79% top-5 vs - BM25 40% / grep 20% on stripped repos; keyword still wins exact - strings). PR attention ≠ correctness - ([egma-ai/jev-reviewer](https://github.com/egma-ai/jev-reviewer); - **not** choxos pointer-not-generator). Skills→oxlint - ([jev-oxlint](https://github.com/cephalization/jev-oxlint); - AST prove ∩ remainder; Phoenix fixtures; not a hard gate; - `tenbin` owns lint). Session-sticky first-prompt routing - ([jev-adaptive-thinking](https://github.com/jxu-dev-c/jev-adaptive-thinking); - fail-closed fallback). Measured RAG rerank - ([Jev-RAG](https://github.com/Max-sm-yc/Jev-RAG); one-run ≥70% - cost / 72% latency vs Spark *rerank*; full-context Spark still - faster). MED: safe-sh, jev-loan-triage, TurboGuo arenas, jevbox; - hermes/mcp packs not found this pass. No wrapper. No invented - metrics. -- Hourly ~19:48 Boise 2026-09-18 / 01:48 UTC 2026-09-19 fold - (`research/notes.md` §59): Archer still Watch. Architecture - notes, not a pip / venv / Cloudflare catalog. Capability kernel - ([interlock](https://github.com/somoore/interlock); LLM ring 3 / - kernel ring 0; secrets never in the agent; Jev SENSOR; - `policy.py` BLOCK/ASK/ALLOW; type-safe ≠ correct; distinct from - toolgate). Typed control plane around DSPy - ([jev-dspy-control-plane](https://github.com/manikanda-kumar/jev-dspy-control-plane); - DSPy drafts AFTER route+action; OpenJEV / DSPy / JSON Schema - share ontology; offline heuristic ≠ quality). Native-probability - calibration arena - ([jev-arena](https://github.com/meetr1912/jev-arena); live 145 - noul Brier 0.0059 / ECE 0.0620 *theirs*; overconfident in low - bins; 2-request fan-out) plus sonar (heatmap-as-policy) / - vickrey (Jev never bids) / bracket (Brier vs Elo; live trailed - Elo). Engine owns truth / Jev owns judgment - ([game-coach](https://github.com/JoelLewis/game-coach); Wave 0 - PRD; Stockfish WASM; GPL-3.0; anti-soundness-theater with egma). - Human-confirmed port cleanup - ([port-cleanup](https://github.com/epiphany-dynamics/port-cleanup); - Jev recommends; human is the only kill trigger; identity - re-check; shields; mapped explanations). MED toolbelt: - jev-pr-labeler, jevcumber, typedecide, jevon, dsh-jev, - fast-jev-compaction-pi, jev-tetris-benchmark, modelsystem, - opencode-system-one, browser-ai, semantic-bookmark. Skip - jef-mcp (parody) and jevregist (account farming). Star spike: - SemIf 1491→1606 (this pass 1607); jevlike 851→896 (this pass - 897). No wrapper. No invented metrics. -- Hourly ~20:43 Boise 2026-09-18 / 02:43 UTC 2026-09-19 fold - (`research/notes.md` §60): Archer still Watch. Architecture - notes, not a uv / bun / Modal catalog. Domain LoRA specialist - vs few-shot hosted - ([Domain-jev-maker](https://github.com/help-er/Domain-jev-maker); - independent CLINC gold, not a Jev teacher-copy; - matched-precision KL 0.168 vs 0.580 banking; few-shot - determinate McNemar n.s.; train when downstream reads p). - Decide→policy→LLM leftover cascade - ([jav-email-cascade](https://github.com/skiingfalcon/jav-email-cascade); - jev vs gen-json vs gen-logprob; Noul 0.5 never rounded; - license null; mock gen-json flat-confidence is *their mock*). - ORDER BY ranking family - ([jev-orderby-bench](https://github.com/yodablocks/jev-orderby-bench); - six gates pass; Score ordinal 0.143 weak link; 53-way 0.99 - tie; calibration ≠ sortable; recodelabs batch-40 fails - ranking). Wire-compat GLiFormer backend - ([jeff](https://github.com/logan-markewich/jeff); typesafe-sdk - drop-in; ~$2.6 vs $15.6 L4 HTTP ~6×; A10G direct ~$0.65 ~24×; - AG News 75.5% vs 90.5%; CPU more expensive; license null; not - a Jev replica). MED: loopback gateway - ([sysone](https://github.com/hraness/sysone); hosted + local - OpenJev/NanoJev/Mini-Jev; no weights; credential from env). - No wrapper. No invented metrics. -- Hourly ~21:39 Boise 2026-09-18 / 03:39 UTC 2026-09-19 fold - (`research/notes.md` §61): Archer still Watch. Architecture - notes, not a pip / npm / bun catalog. Active-learning triage - ([jev-triage](https://github.com/ThyFriendlyFox/jev-triage); - accept / expensive teacher / human; log full distributions; - **do not distill Jev as teacher of record**, ~68% ceiling). - Evidence-packet explorer - ([jev-semantic-explorer](https://github.com/jimmyhealer/jev-semantic-explorer) - / jevex; index-once ask-many; 1/8→6/8 SWE-bench Verified - finish n=8 *theirs*; packet HitFile 0.233 diagnostic). - Meaning-grep - ([jev-semgrep](https://github.com/uehaj/jev-semgrep); AND/OR/NOT - line Nouls; JP↔EN; MIT LICENSE / GitHub NOASSERTION; 0.94/0.98 - *theirs*). Closed-vote CU - ([JevOnly](https://github.com/buluoray/JevOnly); no planner LLM; - 11 steps / 43 calls / ~$0.014 / 17 s *theirs*). Harbor Jev vs - local MLX PCD vs AR JSON - ([system-one-benchmark](https://github.com/mallahyari/system-one-benchmark); - toxic-chat n=50; Jev 84.0% / Brier 0.1096 vs PCD 52% / 0.3884; - O(1) ≠ calibrated Noul; license null). Host-owned product - ([waymode](https://github.com/mossburgh/waymode); app retains - handlers/permissions; 24/26 + 34/36 *theirs*; not a - self-driving proof). OMP/pi fail-open gates - ([omp-jev-extensions](https://github.com/luw2007/omp-jev-extensions); - `jev_acceptance_gate` + `jev_route`; contrast pi-jev-approver - fail-closed). Skip empty jev-compactor / laya-jolt. No wrapper. - No invented metrics. -- Hourly ~22:38 Boise 2026-09-18 / 04:38 UTC 2026-09-19 fold - (`research/notes.md` §62): Archer still Watch. Architecture - notes, not an `omp plugin` / pip catalog. Watch archive path - missing on this VM; receipts from live GitHub. Permission vs - probability - ([omp-greenlight](https://github.com/SemetricLabs/omp-greenlight); - 1,013 calls / 10 sessions; default **40.9%** prompts removed / - **0 of 94** unsafe auto-approvals on labelled corpus; operator - owns thresholds; plugin never self-tunes; not a sandbox; host - deny stays above). Judgment ≠ permission - ([skill-broker](https://github.com/adamjralph/skill-broker); - Hermes pre-agent outline; code owns grants; Jev never grants - access; **not a production recipe**). Eval integrity / - instrument-not-score - ([dinostomp](https://github.com/collapseindex/dinostomp); - FINDINGS 189 / 99 against itself; `dinostomp jev` if-statement - hygiene; ECE 0.062 *theirs* on 24 examples; beside jevals, not - a Harbor taskset). MED: fast-jev-opencode, jev-desktop, - jev-agent-integration, sift, JevExplore. Census: Awesomejev - 488/21644; SemIf 1641 (+13); jevlike 905 (+4); tracker likes - 41 (+1); Laya yes; Blackwood ABSENT; X MCP flapping - (`pages_archived` 0). No wrapper. No invented metrics. -- Hourly ~23:40 Boise 2026-09-18 / 05:40 UTC 2026-09-19 fold - (`research/notes.md` §63): Archer still Watch. Architecture - notes, not a uvicorn / bun / marketplace catalog. Watch - archive path missing on this VM; receipts from live GitHub + - HF. Hunches labeled. Constrained optimizer + S1 features - ([slo-router](https://github.com/zeeshan8281/slo-router); - license null; Jev task/exactness/evidence as features, never - the sole hot-path gate; fail-open local features; same - routes/accuracy; p95 **77.93 → 490.38 ms** *theirs*; eight-row - demo is not a benchmark). Privilege ≠ verdict - ([construct-auto-classifier](https://github.com/godspede/construct-auto-classifier); - Apache-2.0; effect-based shell gate; fast-allow/deny then Jev - Choice + independent risk Nouls; fail-closed; Jev **0** - dangerous / 975; every chat model leaked 16–104; operator-owned - dials). Attention filter / VOI for human review - ([jev-lens](https://github.com/rashedInt32/jev-lens) + - [jev-lens.nvim](https://github.com/rashedInt32/jev-lens.nvim); - never blocks the agent; never edits; never green unless sure). - MED: [sysone-help/sysone](https://github.com/sysone-help/sysone) - (evaluation-model-first TS SDK; **not** hraness/sysone gateway); - [INSTRUCT_JEV](https://huggingface.co/datasets/ctaxnagomi/INSTRUCT_JEV) - (119 rows; 47/51/21; jevals seed); - [swift-jev](https://github.com/ckaik/swift-jev) (LICENSE-only - this pass; not a CLI product). Census: Awesomejev 488/21644; - SemIf **1652** (+11); tracker likes **42** (+1); lastModified - unchanged; Laya yes; Blackwood ABSENT; X MCP flapping. No - wrapper. No invented metrics. -- Hourly ~00:39 Boise 2026-09-19 / 06:39 UTC fold - (`research/notes.md` §64): Archer still Watch. Architecture - notes, not a uvx / pnpm / marketplace catalog. Watch - archive path missing on this VM; receipts from live GitHub. - Hunches labeled. Measurement owns endorsement - ([jev-packs](https://github.com/dtduc-git/jev-packs); - CC0; nine packs `verified` *theirs* on pinned - `jev-1.13.0`; accuracy/ECE/cost/latency; `unknown` - mandatory; named runner jevassert **not released** / 404; - packs without evidence stay `provisional`). Jev supplies - evidence, code owns authority - ([actiongate-jev](https://github.com/omkarghugarkar007/actiongate-jev); - Apache-2.0; deterministic policy owns ALLOW|REVIEW|BLOCK; - positive score never overrides a hard security fail; - fail-closed financial/destructive/credential if Jev is - down; 500-case is label-baseline, not accuracy). Ranking ≠ - calibration - ([does-jev-confidence-mean-anything](https://github.com/Adilmp/does-jev-confidence-mean-anything) - + [jevcal](https://github.com/Adilmp/jevcal); 8,000 - human-annotated judgments; AUC **~0.91**; stated **~75%** - vs human **~10%**; two-parameter recalibration removes - **~96% ECE** without changing rank; never hard-threshold - raw p as a frequency; vendor "calibrated" often means - rank-correlation). MED: - [gqgs/laya-onnx](https://github.com/gqgs/laya-onnx) - (complete Laya→browser int8; distinct from Mattepiu); - [kunchenguid/local-jev](https://github.com/kunchenguid/local-jev) - (ModernBERT local approximation — not equivalence). Do - not re-fold sysone-help/sysone. Census: Awesomejev - 488/21644; SemIf **1660** (+8); jevlike **910** (+5); - TypeAR 9; tracker likes 42; lastModified unchanged; Laya - yes; Blackwood ABSENT; X MCP flapping. No wrapper. No - invented metrics. -- Same-hour remainder ~00:39 Boise 2026-09-19 (`research/notes.md` - §65): Archer still Watch. Do not re-fold actiongate / jev-packs - / sysone-help. Hot-click CU - ([ego-jev](https://github.com/jiangkoumo/ego-jev); MIT; indexed - viewport table → operation+target; code owns observe/execute/ - `--until`; text model only for type; HN 4.9 s vs 9.7 s / wiki - 5.4 s vs 10.1 s *theirs* n=3, high variance, not a bench). - Jev judges relevance, code decides structure - ([jev-compactor](https://github.com/edwardyen724-g/jev-compactor); - MIT; was empty skip §61; never rewrite; regex floor; compaction - fail-open if Jev down, safety fail-closed; 64.5% / 366 ms / - $0.0004 / 0 invented paths / 4 of 4 facts vs Sonnet summary - 96.2% / 1 invented path, one session). Local rules first, - never auto-train on the model's own hides - ([x-reply-filter](https://github.com/zhuyansen/x-reply-filter); - MIT; `rules.js` then batched Nouls; confirm-queue). OpenCode - port already §62: fast-jev-opencode. Census as §64. No wrapper. - No invented metrics. -- Hourly ~01:47 Boise 2026-09-19 (`research/notes.md` §66): Archer - still Watch. Three clusters: **control-plane combinators** - ([decision-combinators](https://github.com/voidning/decision-combinators); - Then/Gate/Vote/Cascade/Weighted; not literal AND/OR; not chat - turns) + **skill VOI** - ([skillranker](https://github.com/Dicklesworthstone/skillranker); - 52★; two-pass + none-of-these; hook **fail-open** — corrects - §7 fail-closed); **eval integrity without leaderboard theater** - ([jev-capability-atlas](https://github.com/Zaious/jev-capability-atlas) - 10★ receipts, type-safe ≠ correct, axis already §49; - [jev-frontier-100](https://github.com/softpudding/jev-frontier-100) - Jev 77.0% vs Qwen3.5 4B/2048 96.7% / 4B off 56.0%, exploratory; - [jev-ood-calibration](https://github.com/scienthoon/jev-ood-calibration) - 900 tickets ECE 0.107 = 4.4× floor, Choice/Score T~3.3 vs - boolean T 0.66, unknowable priority mean p 0.74); **gate - doctrine clone** - ([turnstile](https://github.com/zyphr-labs/turnstile); Apache-2.0; - policy first, Jev remainder, replay; missing Jev → Review); - **MLX one-pass replica economics** - ([jevmlx](https://github.com/bnsd55/jevmlx); 28★; softmax ≠ - Noul; no local leaderboard yet). MED: invalidate (0 of 157 - false invalidations), jev-intent-review (under construction), - prune-review (22-run cost 1.18% with 305% outlier). Census not - re-derived. No wrapper. No invented metrics. -- Hourly ~02:38 Boise 2026-09-19 (`research/notes.md` §67): Archer - still Watch. Do not re-fold the 01:47 list except sibling - contrast. **TLA+ compose with judgment** - ([jev-labs](https://github.com/copyleftdev/jev-labs); MIT; - never confidently wrong; 1,080 golden 0 wrong *theirs* under - chaos, escalate 5%→18% severe; TLC 1,049,750 states / 0 - errors; synthetic, not clinical). **Advance/coverage ledger** - ([seal](https://github.com/Reasonofmoon/seal); MIT; no seal, - no advance; coverage.path auto|code|human|escalate; mint ≠ - product brain). **skill-broker sibling** (outline already - §62; grants in code vs turnstile runtime vs skillranker - advisory). **Sureness** - ([how-sure-is-jev](https://github.com/adarc8/how-sure-is-jev); - MIT; Choice confidence = max_prob; 75/25 → 0.5 vs entropy - 0.19). **JevBench v1.1** - ([jevbench](https://github.com/fstandhartinger/jevbench); - MIT; unofficial; Main Score 0.6/0.2/0.2; Jev 1.13.0 **87.6**; - calibration reported not scored). **CI typed gate** - ([ci-gatekeeper-bot-jev](https://github.com/NemanjaManic/ci-gatekeeper-bot-jev); - package.json MIT / GitHub SPDX null; 504–629 ms *theirs*). - **Codex MCP adapter** - ([jev-in-codex](https://github.com/teempai/jev-in-codex); - MIT; ranking unbenchmarked; lexical fallback). Census: - SemIf **1683** (+11); jevlike **923** (+5); tracker likes - **43** (+1); Awesomejev 488/21644 unchanged. No wrapper. No - invented metrics. -- Hourly ~03:38 Boise 2026-09-19 (`research/notes.md` §68): Archer - still Watch. Do not re-fold the 02:38 list except sibling - contrast. **Judgment as attention redirect, not a merge - blocker** - ([jev-preflight](https://github.com/muse0509/jev-preflight); - Go MIT; eight risk axes; assist=one reinspect; fail-open; - uncalibrated 0.85; owner-run Claude Code 2.1.267: no-key - fail-open PASS, key-enabled exactly one continuation). - **Landed-script trust / headless≠auto-approve** - ([construct-auto-classifier](https://github.com/godspede/construct-auto-classifier) - delta; cert still Jev **0** dangerous / 975; $0.047/1k). - **Jev judges relevance; code decides structure** - ([jev-compactor](https://github.com/edwardyen724-g/jev-compactor) - product-arm **73%** / 350 ms / 4 of 4 *theirs*; 30–250× - cheaper than shipped summarizers; §65 64.5% is vs-Sonnet). - **Compress-before-first-send** - ([dizk/jev-lens](https://github.com/dizk/jev-lens); MIT; - 79% fewer tokens / 500 SWE-rebench; post-send prune +17% - cost; distinct from rashedInt32/jev-lens). **tools≠use** - ([jev-carryforward](https://github.com/Dharundp6/jev-carryforward) - 0/4 recall; SessionStart > hoping). **Observational - memory** - ([pi-observational-memory-jev](https://github.com/willfish/pi-observational-memory-jev); - keep/kind verbatim; model-free compact). **Independent - open-Jev class** - ([openvons](https://github.com/genai-craft/openvons); - Apache-2.0 LICENSE / GitHub SPDX NOASSERTION; 7★; JevPick - 3.2–4.8×; `/v1/systemone` wire-compat ≠ replica). - **Physical-world S1** - ([HA-Jev](https://github.com/AboveColin/HA-Jev); MIT; - **17★**; sensors from typed answers; not for - locks/heaters). **Judgment outside the store** - ([jevql](https://github.com/kylemclaren/jevql); CLI - judges; vanilla Postgres never sees `jev()`). Short - consumer bullet: [sift](https://github.com/bohutang/sift) - ~$0.00003/post. Census: Awesomejev **561** (+73, agent - tooling 87→107); SemIf **1704**. No wrapper. No invented - metrics. -- Hourly ~04:39 Boise 2026-09-19 (`research/notes.md` §69): Archer - still Watch. Do not re-fold the 03:38 list except sibling - contrast / combinators rename. **Digital-design combinators** - ([jev-combinators](https://github.com/voidning/jev-combinators) - is the rename of decision-combinators; extended Router / - Loop / Retry / Fallback / Memory; metaphor ≠ literal AND/OR). - **VOI cache admission** - ([jevcache](https://github.com/kushals256/jevcache); MIT; - same-intent skip LLM; n=100 *theirs* 0 FP / precision 1 / - recall 0.38 / fpr 0 vs Jaccard@0.35 fpr 0.48; fail-open). - **Harbor skill-routing harness** - ([pi-jev-skill-bench](https://github.com/iamdin/pi-jev-skill-bench) - + [pi-jev-skill-suggestion](https://github.com/iamdin/pi-jev-skill-suggestion); - BM25 vs Jev at roster 50–500; 43 gold; no live numbers this - pass; no-key no-op; tool mode is tools≠use cousin). - **Zeroshot vs BERT displacement** - ([jev-zeroshot-vs-bert](https://github.com/zhuyansen/jev-zeroshot-vs-bert); - +0.05–+0.13 vs DeBERTa-c; contamination 0.901 vs `-c` 0.763; - ≈230 / >2048 labels; DiD 0.035 vs 0.112 *theirs*). - **Typed escalate/continue/abort baton** - ([jev-handoff](https://github.com/shitianfang/jev-handoff); - MIT; inverted loop; gate never grants; fail-open; Vercel - drops confidence). **Worth-your-attention VOI** - ([ThinkyMiner/Winnow](https://github.com/ThinkyMiner/Winnow); - MIT; 80%/90% *theirs*; **≠** kevinpita/winnow). - **Jev WHETHER / Python HOW / LLM WHAT** - ([hermes-jev-router](https://github.com/rsdkrasen/hermes-jev-router); - license null; community plugin; skip-next needs core patch). - **Conflict ≠ ignorance** - ([jev-typed-evaluation-collapse](https://github.com/mleyvaz/jev-typed-evaluation-collapse); - Noul collapses; named Choice p=1.0; binary red 0.67–0.85 - *theirs*). **Playwright executes, Jev chooses** - ([browser-jev](https://github.com/DowLucas/browser-jev); - license null; sample-from-distribution). **Local class** - ([OpenJev](https://github.com/IamBusy/OpenJev) Apache-2.0 - `/v1/decide` 45/60 *theirs*, not TypeSafe drop-in, ≠ - hraness/sysone runners; - [semif-serve](https://github.com/dddanielliu/semif-serve) - 1164 vs 178 ms; runoff ≠ softmax; wire-compat ≠ replica). - Toolbelt notes: jev-security-scan / jev-decisions / TeoMastro - (summary.md 404 this pass); **rh-guard owns reward-hack**. - Flywheel: - [DGUI_HYPERMEM-JEV](https://huggingface.co/datasets/ctaxnagomi/DGUI_HYPERMEM-JEV) - 6-row schema. Census: Awesomejev **flat 561/27007**; tracker - likes **43→45**; SemIf **1714** (+10); jevlike **926** (+3). - No wrapper. No invented metrics. -- Hourly ~05:46 Boise 2026-09-19 (`research/notes.md` §70): Archer - still Watch. Do not re-fold §50–§69 HIGH except sibling - contrast / jevassert landing / prune-review, intent-review, - laya-jolt, local-jev deltas. **Record/replay CI LANDED** - ([jevassert](https://github.com/dtduc-git/jevassert); - Apache-2.0; accuracy/ECE/Brier/cost/latency offline from - recordings; exit 0/1/2; McNemar; Action `@v0`). - **Evidence-gated packs now have a runner** - ([jev-packs](https://github.com/dtduc-git/jev-packs); - size 0→458; 2,990-case matrix *theirs*: Jev/Sonnet 5 - accuracy tie Δ≤0.018, Jev better calibrated 7/9, ~250× - cheaper; sms-spam this-pass 0.953/ECE 0.040). - **Failure-finding arena** - ([jevarena](https://github.com/chenmingtang830/jevarena); - Apache-2.0; **≠** meetr1912/jev-arena; harness not findings). - **BBQ stereotype/uncertainty/cost** - ([jev-bbq-experiment](https://github.com/simonmesmith/jev-bbq-experiment); - license null; 58,492; 97.28%; bias 0.04/0.34; $0.3429 / - 7.75 min *theirs*; not a general bias cert). - **Decider ≠ executor** - ([jeffrey](https://github.com/thomasbrueggemann/jeffrey); - MIT; Jev next-tool/progress/risk/done; LLM fills args; - pick ≠ fill). **Sentence-as-rule lint** - ([jevlint](https://github.com/mizchi/jevlint); MIT; - ast-grep × `ask:`; 13/15 1.00/1.00 *theirs*; **≠** - huntedman/JevLint). **VOI hunk prune** - ([prune-review](https://github.com/shubhangi013/prune-review); - 22-run 1.18% with 305% outlier; ~20% target; cost not - quality). **Whole-repo intent** - ([jev-intent-review](https://github.com/yottayoshida/jev-intent-review); - VERIFIED/VIOLATION/UNKNOWN; empty search ≠ proof). - **GLiNER2 System One spec** - ([Jev_from_GLiNER2](https://github.com/Eran-BA/Jev_from_GLiNER2); - spec-only; ≠ jeff). **Open replica substrates** - ([grande](https://github.com/bokuweb/grande) JGLUE 0.614/ - 0.853 + 270M 0.710/0.710 *theirs*; - [laya-jolt](https://github.com/jlt-commons/laya-jolt) - byte parity; [JEV-CPU](https://github.com/leesk212/JEV-CPU) - PoC, Meanblock 404; [local-jev](https://github.com/kunchenguid/local-jev) - done 30%/shape 57%). **Persist constraints** - ([pi-heed](https://github.com/Nyarlathoteppppp/pi-heed); - 98.5%/0 false block *theirs*). Toolbelt note: - actiongate slogan already §64. MED: - [system-one-responsible-ai](https://github.com/david-j-lustig/system-one-responsible-ai) - size-0 framing stub. Census not re-derived. No wrapper. - No invented metrics. -- Hourly ~06:43 Boise 2026-09-19 (`research/notes.md` §71): Archer - still Watch. Do not re-fold §50–§70 HIGH except sibling - contrast. **Harbor SGR-judge contract** - ([jev-judge-bench](https://github.com/slavadubrov/jev-judge-bench); - README MIT / GitHub SPDX NOASSERTION; frozen SLA-150; Jev vs - Luna / DeepSeek-flash / glm-5.3-flash; invalid = FN; - 21 offline tests; canaries ≠ quality; **no quality headline - yet**; **≠** jevarena / jevbench). **Empty skip** - ([jev-context-pruner](https://github.com/IPECTER/jev-context-pruner); - 409 empty). **Hand no-text steps** - ([jev-use](https://github.com/shitianfang/jev-use); MIT - v0.4.1; p50 220 ms; 186 vs 2,672 ms; gate 12/12; Vercel - drops confidence → margin 0.4; first loop 17/20 then 0/20 - *theirs*; **≠** jev-ultrafast). **Pi System-One control - plane** ([pi-jev-control](https://github.com/goodruizhan/pi-jev-control); - license null; v0.3.0 private; GUI never force-click; - compaction never writes session). **Never free-generates** - ([jev-gpt](https://github.com/florian-hoenicke/jev-gpt); - license null; ~400 calls / 75 s / 2¢ *theirs*). - **OpenRouter recipe atlas** - ([jev-cookbook](https://github.com/nexibeo/jev-cookbook); - MIT; 1★; 16–36 samples not benches; 425 calls / $0.015; - browser 5/6 *theirs*). **Personal-history feed** - ([jevfeed](https://github.com/fengyiqicoder/jevfeed); MIT; - no social graph; one request per batch of ten). - **Competing NAR claim-audit, not endorsement** - ([openJev-verdict-2.0](https://github.com/Heman10x-NGU/openJev-verdict-2.0); - README Apache-2.0 / GitHub SPDX NOASSERTION; 77.10%/0.0636/ - 0.0144 *theirs* unverified; **open PR #1**: throughput≠ - latency, Laya parity, like-for-like ECE; **≠** - IamBusy/OpenJev). Census not re-derived. No wrapper. No - invented metrics. -- Hourly ~07:49 Boise 2026-09-19 (`research/notes.md` §72): Archer - still Watch. Do not re-fold §50–§71 HIGH except sibling - contrast. **1-token logprob endpoint ≠ Noul** - ([chakuho](https://github.com/taku-me/chakuho); MIT; - coverage ≠ correctness; GUI 336 *theirs* 27B 95%/92% - vs Jev 89%/82%; `__none__` 97% vs 8B 10%). **Open - replica engine** ([jevinf](https://github.com/zerodegress/jevinf); - MIT; 2.57×/2.27× 100% argmax; MPS only). **Unofficial - Elixir SDK ≠ OTP peer** - ([typesafe-elixir-sdk](https://github.com/phiat/typesafe-elixir-sdk); - MIT; 1★; ≠ dannote/jev). **jevex rename + n=16 VOI** - ([jevex](https://github.com/jimmyhealer/jevex); 160s→69s - / $8.74→$3.13 / 16/16 *theirs*; keep n=8 1/8→6/8). - **Commit attention≠verdict** - ([commitjev](https://github.com/yodablocks/commitjev); - MIT; middle band never rounded; 0 false on 5 clean - *theirs*). **Hermes plugin is Agnes not TypeSafe** - ([hermes-plugin-jev](https://github.com/Mrmimee/hermes-plugin-jev)). - **Pi compact ≠ compaction** - ([pi-jev-compact](https://github.com/dev-willbird1936/pi-jev-compact); - MIT). **Empty skip** - ([jev-runway](https://github.com/IPECTER/jev-runway); - LICENSE-only). **Decision-native inbox** - ([mailordinal](https://github.com/Milo318/mailordinal); - MIT). **Unofficial jev-cli not ready** - ([jev-cli](https://github.com/shaharia-lab/jev-cli); - 0.0.0; ≠ jevql). **Laya multilingual** - ([laya-multilingual](https://huggingface.co/convaiinnovations/laya-multilingual); - MASSIVE 0.366/0.387; Khmer 0.000@0.952; ships - uncalibrated). **Schema-scorer Hub** (GitHub 404; v2 - Choice 0.841; peaked ranking). HF 401 this pass on - open-jev-laya-bench / jev-tree-choice-cap / - INSTRUCT_JEV; jevlogs 404+401. Census not re-derived. - No wrapper. No invented metrics. -- User-provided signal ~08:37 Boise 2026-09-19 - (`research/notes.md` §73): **Skip Archer.** - **Productized System One HTTP** - ([classifier-dev](https://github.com/mrmps/classifier-dev); - MIT; **185★**; https://classifier.dev). Label + - calibrated confidence as the public contract; batch - `{id,text}[]` ~1000; Jev primary, LLM fallback only. - 400 headlines **650 ms** *theirs*. **Escalate-under- - threshold:** smart re-asks single-label <0.7; - multi-label ignores (re-judge worse, 23 s). Emotion - ≥0.9 → 82% / <0.5 → 29%; gemini-3.8-flash 87.5→90.0 / - 61.8→63.7 *theirs*. Multi-label F1 **0.887** / **230 ms** - vs cascade **0.799** / 1.5 s (eval 232 ms; AG News - **87.7%** vs 82.0%). **Measurement-first:** `/benchmark` - from tracked JSON; read eval/README (n=7 train-on-test; - ~0.03 coin flip). **Silent FALLBACK:** granite F1 - **0.546** vs advertised ~**0.800** *theirs*; rh-guard - owns the gate. Life/business (spam/inbox/feedback), - not SWE-only. Distinct from ask-jev-ai wall. No wrapper. - No invented metrics. -- User-provided signal ~08:48 Boise 2026-09-19 - (`research/notes.md` §74): **Skip Archer.** **Delta of - §48.** Pointer-not-generator at evidence-synthesis - scale - ([choxos/jev-reviewer](https://github.com/choxos/jev-reviewer); - MIT; **12★**; https://jevreviewer.xera.ac). **≠** - [egma-ai/jev-reviewer](https://github.com/egma-ai/jev-reviewer). - Two-pass Choice (which line) + Noul (does this line - itself answer); quotes = Noul ≥ 0.5 *theirs*. *Not - found* / *Unclear* first-class. Human tick is the - product (checked never overwritten). 18-q template - **4.6 s / $0.0101** *theirs* (spot check, not a - validation study). Cochrane / PRISMA / RoB, not - SWE-only. No wrapper. No invented metrics. -- User-provided signal ~08:56 Boise 2026-09-19 - (`research/notes.md` §75): **Skip Archer.** - **Wire-compat ≠ logit-equiv** - ([githubnext/localjev](https://github.com/githubnext/localjev); - MIT; **261★**; GitHub Next). **≠** - [kunchenguid/local-jev](https://github.com/kunchenguid/local-jev). - Bun `POST /v1/systemone` on DiffusionGemma via Chat - Completions; TypeSafe SDK drop-in. Prompted JSON → - validate/retry → normalize + entropy confidence — not - razorback16 structured-read logits. Harbor-shaped - bake-off *theirs*: 1,200 req; Qwen3.6 short macro - **76.7%**; Gemma 4 26B-A4B **75.0%**; DiffusionGemma - **74.2%**; no definitive winner (2/120); do not treat - as calibrated. LM Studio still cannot load - DiffusionGemma. Do not copy bun / `.env`. No wrapper. - No invented metrics. -- User-provided signal ~09:07 Boise 2026-09-19 - (`research/notes.md` §76): **Skip Archer.** **Laya - packaging, not a new species** - ([NandhaKishorM/laya](https://github.com/NandhaKishorM/laya); - Apache-2.0; **710★**). PyPI + `Router` over Hub - [`laya`](https://huggingface.co/convaiinnovations/laya) / - [`laya-multilingual`](https://huggingface.co/convaiinnovations/laya-multilingual) / - [`laya-typed-decisions`](https://huggingface.co/convaiinnovations/laya-typed-decisions). - **≠** TypeSafe `/v1/systemone`. **≠** githubnext/localjev. - T4 *theirs*: 1q **32.8 ms** (~7.8× vs Jev p50 - 236–276 ms). Post-T ECE **0.081** vs Jev **0.246**; - raw ECE still trails (0.213 vs 0.144). Banking77 - **0.425** vs Jev **0.870** (77 vs 72; ~3–4 tok/label). - typed-decisions **0.766** is a fine-tune (base - 0.362/0.342 vs majority 0.461). Soft-acc 0.471 vs - 0.580. Khmer **0.000@0.952** — Router because gating - cannot catch. 0.85 still soft. Jev rows third-party - unpublished-here. Do not copy pip / preload. No - wrapper. No invented metrics. -- User-provided signal ~09:14 Boise 2026-09-19 - (`research/notes.md` §77): **Skip Archer.** **External - openjev census ≠ scored bake-off** - ([@airesearch12](https://x.com/airesearch12/status/2101259522933186879); - Florian S / Benchmark Heaven). Named ~18 (system-one-open, - openjev-sglang, DeBERTa open-jev, Needle 3, - open-alternative-jev, Nimble 9B, SemIf, open-jev Dasein / - JoshuaSP, OpenJev razorback16, mini-jev, system-one, - system-one-gemma, jevlike, AlexWortega/openjev, GLiNER2, - Succinct Router 14M, jev-model-router/Director/Loki). - GLiNER2 + routers are **class-boundary**. Incomplete vs - Laya / localjev / kev / TypeAR / openvons. Engagement - **ephemeral** (SIGNAL ~417/9/3; this pass 564/15/5). **≠** - jevbench v1.1. Watch - [jev-models](https://benchmarkheaven.com/jev-models); do - not paste live ranks. Do not copy Stripe. No wrapper. No - invented metrics. -- User-provided signal ~09:24 Boise 2026-09-19 - (`research/notes.md` §78): **Skip Archer.** **JevBench v1.2 - scored board** - ([benchmarkheaven.com/jev-models](https://benchmarkheaven.com/jev-models); - harness [fstandhartinger/jevbench](https://github.com/fstandhartinger/jevbench) - MIT; 0★; HEAD `27ed3d6c`). Protocol `jevbench::v1.2`; scored - 19 Sept 2026; 534 decisions (hard 220 = 30% of Intelligence). - Official Score = geometric mean of I/C/S/K at 25% each. Jev - 1.13.0 **75.3**; SemIf (Qwen3.5-4B) **74.6** (−0.7); OpenJev - DiffusionGemma (razorback16) 67.6 *theirs*. Luna Intelligence - **96.8** rank **#7** on cost. Calibration **on** the rank - (delta from v1.1). Weighting is a product design. Option-order - 72%→21%. Self-host latency ×2 is an assumption; many costs - est. Laya absent (gap, not named-excluded). GLiNER2 mapping - issues; apps out. Qwen3.8 27B Chutes TEE **≠** Archer. **≠** - tweet census §77 **≠** v1.1 87.6. Do not copy Stripe / CLI. - No wrapper. No invented metrics. -- Hourly System One watch ~08:42 Boise 2026-09-19 - (`research/notes.md` §79): **Skip Archer.** Named HIGHs - **already folded** (§73–§78) — extract **how-to-apply**, - not a hit list: wire-compat ≠ logit-equiv (prompted JSON - ≠ structured logit); productize label+p and mark - `FALLBACK`; packaging ≠ new species / script-before-p / - 0.85 still soft; pointer-not-generator (two-pass; *Not - found*; human tick); external census ≠ scored bake-off / - geo-mean weights are a design. **Skip thin noise** - (JEValuate / jevspeak / fable-jev; jev-semgrep already - §61). Hard-gating a Noul as a PR/quality gate is - soundness theater - ([totally-tim/jev-gate](https://github.com/totally-tim/jev-gate) - 0★ ≠ jev-gateway; [claude-jev-warden](https://github.com/connectedGraph/claude-jev-warden) - 1★). Qwen3.8 27B ≠ Archer. No wrapper. No invented - metrics. -- User-provided HIGH ~09:50 Boise 2026-09-19 - (`research/notes.md` §80): **Skip Archer.** Delta of - §46, not a new species. - [khordoo/jev-reflex-autonomy-lab](https://github.com/khordoo/jev-reflex-autonomy-lab) - (TypeScript; **7★**; license null; README SHA - `130987c9`; ARCHITECTURE SHA `48da0769`; HEAD - `e3297ebe`). S1 never stalls waiting; S2 is one-use - advisory and never flies. Purple confidence = - **consumed** S2 (purple bar = arrival; red = fail). - Local controller is rule-based **≠** - githubnext/localjev **≠** kunchenguid/local-jev. - Live API `POST /v1/systemone` `jev-latest`; 20% - starting gate *theirs* still soft and does not start - a mission. Seed = geometry ≠ async replay. No pixels - to either provider; confidence ≠ selected - probability; S2 never grants. README GLM 5.3 vs - ARCHITECTURE muse-spark-1.3-contributor — quote both - *theirs*. Experimental viz, not a flight controller. - Do not copy npm / `.dev.vars`. No wrapper. No - invented metrics. -- User-provided HIGH ~09:51 Boise 2026-09-19 - (`research/notes.md` §81): **Skip Archer.** Productized - observe→score-among-candidates→code-acts on a Mac, - not a new species. - [awlevin/typesafe-computer-use](https://github.com/awlevin/typesafe-computer-use) - (Python; MIT; **427★**; README SHA `369f4a6a`; HEAD - `cc7b5066`). OCR+AX → numbered items → TypeSafe - Choices (`kind`/`item`/`site`/`offscreen`) → - deterministic click/type. **Never ships a screenshot - to frontier for the *decision***; the one-shot - **answer** writer may receive the capture (reader - packet, not the Choice). Writer only for free text. - Overlapping options = false low confidence. AX bonus - never sole (Spotify 0 *theirs*). Post-type Noul 0.5 - and `--min-confidence` 0.4 still soft. $0.0002 vs - Opus $0.032 (155×) *theirs* on **one screenshot**, - not a Harbor taskset. Honest caveat: dates.py rebuilds - pixel-free reasoning. **≠** jev-ultrafast **≠** - cua-s1 **≠** jev-macos-loop **≠** camoufox. Do not - copy `uv sync` / `.env`. No wrapper. No invented - metrics. -- User-provided HIGH ~10:01 Boise 2026-09-19 - (`research/notes.md` §82): **Skip Archer.** Productized - ASR observe→score-among-candidates→code-acts in - headed Chromium, not a new species and not omni. - [moritzkremb/jev-voice-browser](https://github.com/moritzkremb/jev-voice-browser) - (JavaScript; MIT; **103★**; README SHA `fa033303`; - HEAD `054db0f3`). Web Speech partials → one 9–11- - question Jev request (~250–350 ms *theirs*) → - policy. Pointer-not-generator for spans. Closed-set - may act on a partial; free-text waits. Spoken - confirm is convenience, not auth. Numbered overlay, - no second model. Integration 27/27 / ~$0.0002/call - *theirs* fixtures, not a Harbor taskset. 0.5 / 0.55 - / 0.6 still soft. **≠** jev-voice-control **≠** - nikolas-j **≠** Aj1905 **≠** typesafe-computer-use - OCR. Do not copy `npm` / `.env` / `run.sh`. No - wrapper. No invented metrics. Expand the §39 tweet; - do not re-card it. -- User-provided HIGH ~10:20 Boise 2026-09-19 - (`research/notes.md` §83–§84): **Skip Archer.** Two - signals, one fold. - [reddpy/AgentGhost](https://github.com/reddpy/AgentGhost) - (TypeScript; MIT; **2★**; README SHA `44145fa9`; - HEAD `ac04e4fb`). Intent-aware ALLOW/ASK/DENY - wrap-as-execution: the wrap *is* the tool function; - rules first; ASK throws; `failMode: closed`. Judge - is a slot. Provider-hosted tools out of reach. - **≠** jwen5419807/agentghost **≠** vventirozos - **≠** actiongate **≠** toolgate **≠** jev-use. - rh-guard owns the gate cousin. Do not copy `npm` / - `.env` / `AUTO_APPROVE`. [@studio_yebisu JP genre - atlas](https://x.com/studio_yebisu/status/2101065176069886152) - (2026-09-18T21:45:48Z). Apps by hole, not a scored - bake-off. Stars research-time (typesafe-computer-use - 203→**427**; jev-voice-browser 40→**103**). Not - verified evals. Engagement ephemeral (this pass - 131,234 / 1,934 / 192). SAM 3.1 already §39. - OpenRouter Jev no-waitlist is WATCH, not a recipe. - **≠** @airesearch12 class census **≠** v1.2 board. - Do not dump the 30 repos. No wrapper. No invented - metrics. -- User-provided HIGH ~10:25 Boise 2026-09-19 - (`research/notes.md` §85): **Skip Archer.** External - pedagogy / how-to-apply, not a new species. - [@akshay_pachaar “Jev Clearly Explained”](https://x.com/akshay_pachaar/status/2101037514945597645) - (article https://x.com/i/article/2100940576741093376; - 2026-09-18T19:55:53Z). LLM hammer for bounded - decisions; code owns branches; parallel questions; - thresholds in code; **schema-safe ≠ correct**; - placements = routing / tool-risk / verify with LLM; - shadow-mode; questions-as-code. **200× / 400×** and - 70–500 ms / $0.042/MTok are TypeSafe **ceiling** - claims *theirs*, not Harbor. Text-only; not looking - at the screen. **≠** official docs **≠** Flavio - Copes **≠** LangChain harness **≠** AgentGhost. - Engagement ephemeral (this pass 233,495 / 2,280 / - 235). Do not copy the Python samples. No wrapper. - No invented metrics. -- User-provided HIGH ~10:30 Boise 2026-09-19 - (`research/notes.md` §86): **Skip Archer.** Dedicated - fold of [uehaj/jev-semgrep](https://github.com/uehaj/jev-semgrep) - (light-noted §61). Grep by meaning via Jev Noul; - proposition ≠ embedding; contrast-set (all six - about a refund; only customer-*asking* pass); - AND/OR/NOT are boolean ops on *thresholded* bits - (do not multiply p; ≠ jev-combinators metaphor). - Cross-lingual; no index; EN safer near threshold. - Semgrep.dev SAST name collision. **Not a gate** - (ranking fail-open; rh-guard skip). LICENSE MIT / - GitHub NOASSERTION. HEAD `21120e9`; README SHA - `923e6a5`. Stars ephemeral (0 → SIGNAL ★42 → **51** - this pass). 0.94/0.98 LLM-as-judge 10×51 *theirs*, - not Harbor. **≠** jevgrep **≠** jev-sift **≠** - jevex **≠** semgrep.dev. Do not copy npm / `npx` / - `.env` / marketplace. No wrapper. No invented - metrics. -- Hourly 1047 HIGH + deferred 0945 backlog - (`research/notes.md` §87): **Skip Archer.** Docs-only - off main (PR #2 merged). How-to-apply / mental - models / architecture / Harbor-jevals — not - SWE-only. Formal methods compose with scoring; a - Noul is a SENSOR; hard-gating as test/PR/HA - write/authorship is soundness theater. Ten - clusters: decision-validated UI - ([gram-render](https://github.com/wei-b0/gram-render) - never authors text; - [jev2ui](https://github.com/dglazkov/jev2ui) Jev - decides / Gemini writes); - decision-as-assert - ([jevtest](https://github.com/realZachi/jevtest) - ambiguous band 0.15–0.85; 0.85 still soft); - hybrid S1 - ([anima3](https://github.com/hulryung-uo/anima3) - Qwen logprob default; jeff confidently flat; do - **not** invent Laya); - pointer search - ([JevFind](https://github.com/Peu77/JevFind)); - Harbor trio - ([jev-frontier-bench](https://github.com/manjunathshiva/jev-frontier-bench) - 72.5%/ECE 0.161 vs Fable 84%/0.064 *theirs*; - ChaosNLI JS worse than uniform; **≠** - frontier-100; - [jev-gliclass-bench](https://github.com/JoeSlain/jev-gliclass-bench) - product bakeoff 78/40/49; - [job-posting-triage](https://github.com/geckguy/job-posting-triage) - majority floor 0.947 / tfidf wins / calibration ≠ - discrimination); - authorship named escape (not evidence); - non-SWE - ([ha-switchboard](https://github.com/grayslawson/ha-switchboard) - HA remains execution **≠** HA-Jev; - [n8n-nodes-jev](https://github.com/vibe-with-me-tools/n8n-nodes-jev) - unofficial Low Confidence); - compaction delta - ([fast-jev-compaction-pi](https://github.com/zaycruz/fast-jev-compaction-pi) - **≠** pi-jev-compact **≠** pi-jev-compaction; - ~50× *theirs*); - full-distribution optimizer - ([jevloop](https://huggingface.co/spaces/async-dime/jevloop) - UCB1+CEM; no LLM in the loop; mock default); - deferred class - ([laya-vision](https://huggingface.co/thaitea/laya-vision-smolvlm-256m) - SmolVLM; `score` untrained; **≠** blackwood **≠** - Archer; - [Cerebellum-2B](https://github.com/mkeco/Cerebellum-2B) - `/v1/decide` ≠ TypeSafe; wire-compat vs - agent-routing as separate Harbor axes; competing - NAR **not endorsement**; - [laya-grounded](https://huggingface.co/Luni/laya-grounded) - not drop-in; phishing 0.611→0.512; Platt not - temperature). 0★ HIGHs still get real cards. Do - not copy npm / `pi install` / n8n / HA add-on / - `TYPESAFE_API_KEY`. No wrapper. No invented - metrics. -- Queued user SIGNALs + remaining deferred 0945 HIGH - (`research/notes.md` §88): **Skip Archer.** Docs-only - on PR #3. Mental models: open LoRA replica - ([GestaltLabs/Jeff-1](https://huggingface.co/GestaltLabs/Jeff-1) - acc **0.8183** ECE **0.0807** vs Jev **0.8283** / - **0.0932** n=9730 *theirs*; set reused; **≠** - [logan-markewich/jeff](https://github.com/logan-markewich/jeff)); - Jev-first bounded agent - ([stanley-code](https://github.com/devagrawal09/stanley-code) - empty findings ≠ approval; human promote; 0.6/0.55/0.15 - still soft); - NL memory → beam-search FS - ([findme](https://github.com/marc2332/findme) **≠** - JevFind); - price workers not the conversation - ([jevsubrouter](https://github.com/leftspace89/jevsubrouter) - fail-open; counts ≠ dollars). laya-vision + - Cerebellum already §87 — not re-carded. Soft Noul ≠ - hard safety. Do not copy `uv` / npm / cargo / - marketplace / `TYPESAFE_API_KEY` / `JEVSUB_API_KEY`. - No wrapper. No invented metrics. -- Hourly 1144 HIGH (`research/notes.md` §89): **Skip - Archer.** Docs-only on PR #3. Do **not** re-fold - 1047 / §87 / §88. How-to-apply / mental models / - architecture / Harbor-jevals — not a thin Jev skill - dump. Backend-agnostic categorization/scoring/ - decision class. Formal methods compose with scoring; - a Noul is a SENSOR; hard-gating a default 0.5 bool, - quoting apa “mathematically fulfilled,” treating A/B - proxies as token savings, letting Jev send mail, or - pasting jev-test bars as results is soundness theater. - Seven clusters: **Typed if** - ([feelings](https://github.com/BoundaryML/feelings) - `.feels()` default 0.5 is Noul-0.5-never-rounded; - exhaustive BAML `match`; **≠** hunch **≠** Probably; - license null; **0★**); - **Shadow then honor** - ([apa-agent-harness](https://github.com/AiPersonacademy/apa-agent-harness) - **≠** AntonioCoppe/jev-harness; unpublished npm; - 0.85 still soft; - [apa-persona-engine](https://github.com/AiPersonacademy/apa-persona-engine) - SM then leftover LLM; <250 ms ≠ microsecond; - [grok-bot-jev](https://github.com/Bodila51/grok-bot-jev) - skill honor; A/B proxies ≠ tokens; 13.0× is a - top-five cap); - **Human every action** - ([Essentiel-Jev](https://github.com/JacquesGariepy/Essentiel-Jev) - never authority; 0.75 provisional; license null); - **Atom then sense** - ([enzo-mcp](https://github.com/mahawi1992/enzo-mcp) - independently falsifiable claims; UNKNOWN useful; - **≠** jev-sift); - **File by Choice** - ([pigeonhole](https://github.com/noripto/pigeonhole) - OTHER skip; 0.6 still soft; **≠** jev-semgrep; - client-side playground - [jev-agent-decision-playground](https://huggingface.co/spaces/bojansandhaus/jev-agent-decision-playground) - static no-network; **≠** classifier.dev; sibling - jev-decisions pointer only); - **Question preflight** - ([jev-reliability](https://github.com/vcjdeboer/jev-reliability) - Nothing about accuracy; noul-gate 0.0%/12.5%/3.6% - *theirs*; **≠** dinostomp; - [clduab11/jev-test](https://github.com/clduab11/jev-test) - bars ≠ scores; **≠** realZachi/jevtest; - [jev-rag-benchmark](https://github.com/erendikmenn/jev-rag-benchmark) - “Jev wins” is not an assumption; **≠** Jev-RAG; - [dairui1/jev-lab](https://github.com/dairui1/jev-lab) - urgent 91% vs Haiku 79% *theirs* synthetic; **≠** - BrendanH18/jev-lab; do not re-card jev-desktop); - **Inbox read-only vs write** - ([jevmail](https://github.com/fazlerocks/jevmail) - `gmail.readonly` ~3¢/1k *theirs*; **3★**; - [mailjay](https://github.com/secondfret/mailjay) - archive/trash after review; license null; **≠** - mailordinal). Soft Noul ≠ hard safety. 0★ HIGHs - still get real cards. Do not copy unpublished npm - `@aipersona/…` / `uv` / `baml toolchain` / Gateway - keys / `TYPESAFE_API_KEY`. No wrapper. No invented - metrics. -- Hourly 1241 HIGH (`research/notes.md` §90): **Skip - Archer.** Docs-only on a **fresh PR off main**. Never - reopen merged #3 / #4 / #5. Do **not** re-fold 1144 / - §89. How-to-apply / mental models / architecture / - Harbor-jevals / toolbelt — not a thin Jev skill dump. - Backend-agnostic categorization/scoring/decision class - (Jev-like speed/econ). Formal methods compose with - scoring; a Noul is a SENSOR; hard-gating AUTO_ACT, - treating ranking logits as frequencies, pasting - 95.2% / 83% plumbing / 36/120 NL2Bash as class - ceilings, or letting Jev send/delete/close pinned - tabs is soundness theater. Twelve clusters: - **Observe→score→act namesake** - ([ZHUBoer/ego-jev](https://github.com/ZHUBoer/ego-jev) - reserved `__none__`; runWorkflow completed ≠ success; - **≠** jiangkoumo/ego-jev; **0★**); - **Decision-as-ranking** - ([jsort](https://github.com/keltokhy/jsort) scores are - relative; Noul not Choice for scale; CommonLit - r=0.824 / ρ=0.841 *theirs*; **1★**); - **Native vs schema-guided Harbor** - ([groundedness-judge-bench](https://github.com/slavadubrov/groundedness-judge-bench) - native vs schema-guided; implicit_true included - in yes; Jev 0.6667 vs GLM 0.7661 *theirs*; - LICENSE MIT / SPDX NOASSERTION; **0★**; **≠** - jev-judge-bench); - **0 promotions / authored vs real** - ([jev_playground](https://github.com/JYeswak/jev_playground) - 0 promotions; routing-backtest 0.0447%; **0★**; **≠** - HF playground); - **Offload + classifier-not-generator** - ([yuyang2230/jev-agent-skill](https://github.com/yuyang2230/jev-agent-skill) - jev-1.13-free; **≠** GodsBoy; **0★**; - [jev-techstack-classifier](https://github.com/swap-mitra/jev-techstack-classifier) - stack_config.json only; license null; **0★**); - **Collapse late** - ([s1_ruby](https://github.com/innocentdiaz/s1_ruby) - collapse late; `undecided?` abstain; **≠** hunch **≠** - feelings; **1★**); - **Unofficial toolbelt** - ([2389-research/judgement](https://github.com/2389-research/judgement) - license null; confidence ≠ winner p; **0★**; - [typesafeai-sdk-rust-community](https://github.com/community-ports/typesafeai-sdk-rust-community) - typesafeai-sdk-community not a new species; **0★**); - **Pointer shell** - ([tpellet/hunch](https://github.com/tpellet/hunch) - exit 3; never-execute list; **≠** carldaws/hunch; - **0★**); - **Preview-first VOI / rubric rewrite** - ([jev-file-search](https://github.com/emilwagman/jev-file-search) - scores not calibrated accuracy; **0★**; - [jev-linkmap](https://github.com/stas4000/jev-linkmap) - Jev never sees S2 prose; LICENSE MIT / SPDX - NOASSERTION; **0★**); - **Life fail-open covers** - ([jev-mail](https://github.com/muhammedilyasy/jev-mail) - metadata only; [tidy](https://github.com/MANISH007700/tidy) - none-of-folders stay; [tab-bouncer](https://github.com/MANISH007700/tab-bouncer) - pinned/audio/current never closed; - [lkclean](https://github.com/stefw/lkclean) Show - fail-open; [jev-yt-time-saver](https://github.com/jaibhasin/jev-yt-time-saver) - Show anyway); - **S1 decide / S2 plan** - ([ORIGIN-CIVILIZATION](https://github.com/JacquesGariepy/ORIGIN-CIVILIZATION) - pause-if-no-Jev; validResponse sums-to-1; **≠** - Essentiel-Jev; **1★**); - **Seed/expand/judge/verify + local daemon ≠ Jev** - ([jev-crawlers](https://github.com/russfranky/jev-crawlers) - risk bands never raw boolean; **0★**; - [jevbrain](https://github.com/Synxneuos/jevbrain) - AUTO_ACT is not a Noul; license null; **9★**). Soft - Noul ≠ hard safety. 0★ HIGHs still get real cards. - Census (user-provided; not re-derived): Archer still - NOT landed (HF empty; tracker likes 49 lastModified - 2026-09-19T18:37:18Z still promised); Laya yes; - Blackwood ABSENT; SemIf 1846 (+17); jevlike 962 (+3); - TypeAR-AI/TypeAR 10 (+1); Awesomejev flat 561/27007. - Do not copy `TYPESAFE_API_KEY` / `ZEN_API_KEY` / - `AI_GATEWAY` / `uv` / `cargo` / wrangler / chrome - unpacked / OAuth client ids / shop URLs / `.env`. - No wrapper. No invented metrics. -- Hourly 1347 HIGH (`research/notes.md` §91): **Skip - Archer.** Docs-only on a **fresh PR off main**. Never - reopen merged #3 / #4 / #5 / #7. Do **not** re-fold - 1241 / §90. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class (Jev-like - speed/econ). Formal methods compose with scoring; a - Noul is a SENSOR; hard-gating argmax as safety, - pasting 97.7% n=130 / 0 hallucination / 8,026 tokens - as class ceilings, or treating a skill named System - One as a judge is soundness theater. Nine clusters: - **Judge harness as control API** - ([judgekit](https://github.com/lexingtonhibiki/judgekit) - YAML classify/score/route/verify; 97.7% n=130 - *theirs*; **0★**; - [typed-judge-kit](https://github.com/Ascurse/typed-judge-kit) - verdict-in-code; MIN_LABELS=20; **0★**); - **Batch packing VOI** - ([decide](https://github.com/alsoleg89/decide) - packing VOI; 0.8 ≠ 80% accuracy; license null; - **0★**; **≠** jev-sift); - **Calibration as product** - ([Jev-Calibration](https://github.com/AnthusAI/Jev-Calibration) - Platt ECE 0.117→0.052; license null; **0★**; - [jev-calibration-arena](https://github.com/pmcclelland/jev-calibration-arena) - never acts; size 0; **0★**; **≠** jev-arena); - **Decision-as-Plugin** - ([openrouter-jev-mcp](https://github.com/ctmx/openrouter-jev-mcp); - [typesafe-mcp](https://github.com/cyrusasco/typesafe-mcp) - noul deadband 0.35–0.65; - [FrancoisChastel/jev-code](https://github.com/FrancoisChastel/jev-code) - ≠ npm jev-code; **1★**; - [claudecode-jev-marketplace](https://github.com/skylence-org/claudecode-jev-marketplace) - fail-open not hot path; - [mcp_jev](https://github.com/pedroknigge/mcp_jev) - packs not ask_jev; - [jev-skill](https://github.com/codaaiteam/jev-skill) - jevtypesafeai.com ≠ TypeSafe); - **Policy-constrained skill select** - ([hermes-switchyard](https://github.com/bgrablin/hermes-switchyard) - ≠ hermes-jev-router ≠ hermes-plugin-jev; **0★**); - **Tiny local econ pruner** - ([nanoprune](https://github.com/dmdjr1409/nanoprune) - 2.8MB ECE 2.58%; 0 hallucination theater; **0★**); - **Observe→score→act cousins** - ([jev-browser-agent](https://github.com/smartdio/jev-browser-agent) - ≠ ZHUBoer/ego-jev; - [omp-jev-web](https://github.com/Dakai/omp-jev-web) - DONE ≠ proof; - [hari007sh/jev](https://github.com/hari007sh/jev) - ≠ dannote/jev; license null); - **Deterministic verify ≠ System One** - ([system-one-skills](https://github.com/0thernet/system-one-skills) - deterministic verify; **0★**); - **Soft-score vs hard-argmax** - ([typed-gate](https://github.com/harshpuri84/typed-gate) - band [0.40,0.60] is refusal; - [pi-jev-gate](https://github.com/fivethirty/pi-jev-gate) - fail-closed; choice is the verdict; rh-guard owns). Soft - Noul ≠ hard safety. 0★ HIGHs still get real cards. - Census (user-provided; not re-derived): Archer still - NOT landed. Do not copy `TYPESAFE_API_KEY` / - `OPENROUTER_API_KEY` / `JEV_API_KEY` / `uv` / `npx` - / plugin-marketplace install / `.env`. No wrapper. - No invented metrics. -- Hourly 1441 HIGH (`research/notes.md` §92): **Skip - Archer.** Docs-only on a **fresh PR off main**. Never - reopen merged #3 / #4 / #5 / #7 / **#8**. Do **not** - re-fold 1347 / §91. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class (Jev-like - speed/econ). rh-guard owns the gate cousins; - Augustus owns placement. Formal methods compose with - scoring; a Noul is a SENSOR; pasting Foq 100%/ECE - 0.2% / Reranker 0.1667 / 400 ms / akpsahan vs-Jev as - class ceilings, or treating Qwen3.8-27B as Archer, is - soundness theater. Nine clusters: - **Self-hosted econ** - ([Foq](https://github.com/yohanargentina-oss/Foq) - ~25ms/2.2GB local; **1★**; - [rev](https://github.com/jaswanthsanjay88/rev) - prefill-only + HF jev-0.5b; **0★**; - [robfrase/jev](https://github.com/robfrase/jev) - planning memo); - **Soft-judgment gate integrity** - ([typesafe_agent_gates](https://github.com/ThiagaoBR/typesafe_agent_gates) - 27/27 / 31/31; [safe-sh](https://github.com/EpicEric/safe-sh) - static remainder; **1★**; - [jev-pastepilot](https://github.com/buberlo/jev-pastepilot) - Confirm before act; rh-guard owns); - **Retrieval as calibrated decision space** - ([Jev-Reranker](https://github.com/uspraveen/Jev-Reranker) - live Jev not yet measured; - [sessionwise](https://github.com/Nasrallah-AL/sessionwise) - opt-in relevance; - [jev-search](https://github.com/savka777/jev-search) - pointer sieve; **≠** kazuhideoki/jev-search - **≠** superagents-lab/jev-search); - **Enterprise reflexes** - ([400ms-agentic-sf](https://github.com/furuCRM-Inc/400ms-agentic-sf) - Salesforce WebMCP; - [typesafe-scheduler-diagnostics](https://github.com/thevilledev/typesafe-scheduler-diagnostics) - advisory); - **Screenshot-free / CU** - ([droidjev](https://github.com/mkruglikov/droidjev) - screenshot-free; - [jevcu](https://github.com/Tewoto1/Computer-use-and-control-with-Jev) - planner still writes); - **Hybrid S1/S2** - ([ha-conversation-jev](https://github.com/luxus/ha-conversation-jev) - Jev→Grok; **1★**; - [dsh-jev](https://github.com/buberlo/dsh-jev) - can only gate; **2★**); - **Harbor-jevals / SRE** - ([jev-classification-benchmark](https://github.com/rachit-srivastava-devx/jev-classification-benchmark) - specified not run; - [jev-luna-pagerduty-trigger](https://huggingface.co/datasets/reachjalil/jev-luna-pagerduty-trigger) - p≥0.50); - **Laya densifies** - ([meldecision](https://github.com/meldltd/meldecision) - laya-go ONNX; - [laya-doom](https://github.com/shantanugoel/laya-doom) - never pixels; - [laya-api](https://github.com/logixism/laya-api) - empty README; - [akpsahan/laya](https://huggingface.co/akpsahan/laya) - ≠ Archer); - **Demos / unofficial toolbelt** - ([jevchess](https://github.com/choxos/jevchess) - engine owns truth; **1★**; - [jev-drive](https://github.com/vedssharma/jev-drive) - sim not AV; - [story-arc](https://github.com/amali-s/story-arc) - Jev never authors; - [jev-hs-assistant](https://github.com/newbie1668/jev-hs-assistant) - HS6; - [jev-plays-starcraft-2](https://github.com/golergka/jev-plays-starcraft-2) - UI-verified ≠ API Victory; **1★**; - [awesome-jev-use-cases](https://github.com/walidboulanouar/awesome-jev-use-cases) - catalog; **2★**; - [typesafe-go](https://github.com/Nibir1/typesafe-go) - ≠ official). Soft Noul ≠ hard safety. 0★ HIGHs - still get real cards. Census (user-provided; not - re-derived): Archer still NOT landed; tracker likes - **50** lastModified UNCHANGED - 2026-09-19T18:37:18Z; SemIf 1873 (+7); jevlike 969 - (+2); TypeAR 10 flat; Awesomejev 561/27007 flat. - Do not copy `TYPESAFE_API_KEY` / OAuth `client_id` / - `uv` / `npx` / `.env`. No wrapper. No invented - metrics. -- SIGNAL fold jevcache + jev-align (`research/notes.md` - §93): **Skip Archer.** Docs-only on a **fresh PR - off main** after #9 merge `059f3670`. **Never - reopen** merged #3 / #4 / #5 / #7 / #8 / **#9**. - Do **not** re-fold 1441 / §92. How-to-apply / - mental models / architecture / Harbor-jevals / - toolbelt — not a thin Jev skill dump. - Backend-agnostic categorization/scoring/decision - class (Jev-like speed/econ). rh-guard owns - HIT-as-truth and training-score auto-accept as - gate cousins; Augustus owns placement. Formal - methods compose with scoring; a HIT and a - training score are SENSOR. Two clusters: - **Decision ledger / memoization** - ([hyperspaceai/jevcache](https://github.com/hyperspaceai/jevcache); - license **null**; **8★** this pass, SIGNAL ★6; - HEAD `a211d13`; README SHA `7c2abe99`; - fingerprint after redact; recall vs decide; - publish fingerprints+answers; CI replay as Harbor - cousin; Cache hit ≠ correctness; - hyperspaceai/jevcache ≠ kushals256/jevcache; - memoize typed decisions; VOI of cache hit); - **GEPA alignment loop** - ([sutro-sh/jev-align](https://github.com/sutro-sh/jev-align); - Apache-2.0; **60★** this pass, SIGNAL ★56; forks - **7**; HEAD `49753df`; README SHA `363fccb7`; - human labels only; score never auto-accepts; - production capture flywheel; sutro-sh/jev-align ≠ - caiovicentino/jev-align; GEPA + System One). Soft - Noul ≠ hard safety. Census (user-provided; not - re-derived): Archer still NOT landed. Do not copy - `TYPESAFE_API_KEY` / `uv` / `npx` / `curl | sh` / - `.env`. No wrapper. No invented metrics. -- SIGNAL fold enzyme + JA ModernBERT + Gemma/Nemotron/djev-dev/Laya - essay (`research/notes.md` §94): **Skip Archer.** Docs-only - on a **fresh PR off main** after #10 merge `30438eff`. - **Never reopen** merged #3 / #4 / #5 / #7 / #8 / #9 / - **#10**. Do **not** re-fold jevcache/jev-align / §93. - How-to-apply / mental models / architecture / - Harbor-jevals / toolbelt — not a thin Jev skill dump. - Backend-agnostic categorization/scoring/decision - class (Jev-like speed/econ). rh-guard owns - guidance-as-hook, unofficial-local-as-Jev, - hosted-bootstrap silent FALLBACK, - LFM-default-as-JA-softmax, Nemotron - “not calibrated replacement”, and Laya - confidence-without-competence as gate cousins; - Augustus owns placement. Formal methods compose - with scoring; `when asked`, unofficial local p, - Nemotron p, and Laya 0.85 are SENSOR. Three - clusters: **Compile-time System One / - questions-as-index** - ([byenzyme/enzyme](https://github.com/byenzyme/enzyme); - license **null**; **63★** this pass, SIGNAL ★62; - HEAD `c91d6b5`; README SHA `9af7c570`; - guidance ≠ hook; catalysts ≠ summaries; - compile-time System One; hosted bootstrap ≠ - silent TypeSafe; ~350×/1000× *theirs* - (GitHub description); **≠** enzymejs/enzyme); - **Unofficial JA ModernBERT cross-encoder** - ([argos1111/modernbert-ja-310m-jev](https://huggingface.co/argos1111/modernbert-ja-310m-jev); - CC-BY-SA-4.0; **2 likes**; sha `07cda235`; - unofficial ≠ TypeSafe; format_version - modernbert-jev/1; Argos1111/jev_local ≠ - us/jev-local ≠ kunchenguid/local-jev; LFM - default ≠ ModernBERT backend; JGLUE JNLI - 92.62% / JComQA 92.40% *theirs*); - **NAR class legitimacy / multimodal / - Router-OOD** - ([@googlegemma](https://x.com/googlegemma/status/2101069861598482817) - ~0.2s *theirs*; - [pst2154/Nemotron_Jev](https://github.com/pst2154/Nemotron_Jev) - **6★**; README SHA `f2f3d052`; HEAD `983cc29` on - `feat/nemotron-decision-lab`; Nemotron ≠ TypeSafe - Jev; not a calibrated replacement; - [Davipar/djev-dev](https://github.com/Davipar/djev-dev) - Apache-2.0; **2★**; README SHA `6d59d020`; - djev-dev complements djev-spark; images as Choice - options; - [Laya essay](https://laya.convaiinnovations.com/); - Laya essay numbers *theirs*; Router/OOD - confidence). Soft Noul ≠ hard safety. Census - (user-provided; not re-derived): Archer still NOT - landed. Do not copy `TYPESAFE_API_KEY` / - `ENZYME_JEV_MODEL` / `curl | bash` / docker / - `pip install laya`. No wrapper. No invented - metrics. -- Hourly 1541 HIGH (`research/notes.md` §95): **Skip - Archer.** Docs-only on a **fresh PR off main**. Never - reopen merged #7 / **#8** / **#9** / **#10** / - **#12**. Do **not** push onto §93 / §94. Do **not** re-fold - 1441 / §92. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class (Jev-like - speed/econ). rh-guard owns the injection-firewall / - CI-gate cousins; Augustus owns placement. Formal - methods compose with scoring; a Noul is a SENSOR; - pasting orchestrator 0.95 / skip_below 0.05 / - OpenRoboto $ as class ceilings, inventing - one-dollar-tahoe ASR/FPR, or treating numbered-choice - softmax as a Noul is soundness theater. Six clusters: - **Decision-as-plugin for SWE** - ([petercr/jev-orchestrator](https://github.com/petercr/jev-orchestrator) - difficulty + policy thresholds + JSONL trace; **0★**; - [Charlyhno-eng/jev-codex-pilot](https://github.com/Charlyhno-eng/jev-codex-pilot) - model + reasoning depth; **0★**; - [SunnyKikiHK/jev-replacement](https://github.com/SunnyKikiHK/jev-replacement) - keep/shadow/hybrid/reject; license null; **0★**); - **Evidence projection** - ([jackboykin/quarry](https://github.com/jackboykin/quarry) - quarry evidence projection; Go MIT; **0★**; **master**); - **Soft judgment integrity** - ([seb4ez/jevguard](https://github.com/seb4ez/jevguard) - calibrator/cache/escape; **0★**; - [guilhem/jev-ci-selector](https://github.com/guilhem/jev-ci-selector) - CI shadow mode; license null; **0★**; rh-guard owns); - **Physical/control first-class domain** - ([Frank-ZY-Dou/awesome-jev](https://github.com/Frank-ZY-Dou/awesome-jev) - Frank-ZY-Dou/awesome-jev robotics/3D/control; license - null; **0★**; text-state, not pixels); - **Harbor-jevals / injection-firewall** - ([PavitarSinghArneja/one-dollar-tahoe](https://github.com/PavitarSinghArneja/one-dollar-tahoe) - TypeSafe Jev defense eval; **0★**; ~74 demo; README - has no ASR/FPR; rh-guard owns); - **llama.cpp replica** - ([webNeat/llama-jev](https://github.com/webNeat/llama-jev) - llama.cpp replica; license null; **0★**; softmax ≠ - Noul). Soft Noul ≠ hard safety. 0★ HIGHs still get - real cards. Census **not provided this hour** (not - re-derived). Archer still NOT landed. Do not copy - `TYPESAFE_API_KEY` / `AI_GATEWAY_API_KEY` / - `EXA_API_KEY` / `GROQ_API_KEY` / `uv` / `npx` / - `go install` / `.env` / `attacks.json`. No wrapper. - No invented metrics. -- Hourly 1639 HIGH (`research/notes.md` §96): **Skip - Archer.** Docs-only on a **fresh PR off main**. Never - reopen merged #7–**#13**. Do **not** re-fold 1541 / - §95. How-to-apply / mental models / architecture / - Harbor-jevals / toolbelt — not a thin Jev skill dump. - Backend-agnostic categorization/scoring/decision - class. Formal methods compose with scoring; a Noul - is a SENSOR; treating zen-chat as calibrated Jev, - pasting tamaratran 24/24 onto this host, or - hard-gating keepThreshold 0.5 as proof of - irrelevance is soundness theater. One HIGH cluster: - **OpenCode host-port of evidence-preserving stdout - prune** - ([indiejoseph/opencode-jev-pruner](https://github.com/indiejoseph/opencode-jev-pruner) - OpenCode jev-pruner context sieve; - observe→score-candidates→prune; jev-zen / - jev-1.13-free; zen-chat ≠ Noul; fail-open original; - keepScore >0.1 floor; GitHub license **null**; - `package.json` MIT; **0★**; HEAD `764169c` - **master**). MEDIUM watch (do not over-weight): - [jongyunhur/jev-webagent-bench](https://github.com/jongyunhur/jev-webagent-bench) - empty stub; [Kiln-AI/jev_jsonschema](https://github.com/Kiln-AI/jev_jsonschema) - noul_threshold 0.5 (**5★**); [NSStudent/JevSwiftSDK](https://github.com/NSStudent/JevSwiftSDK) - unofficial (**5★**). Soft Noul ≠ hard safety. Census - **not provided this hour** (not re-derived). Archer - still NOT landed. Do not copy `TYPESAFE_API_KEY` / - `OPENCODE_API_KEY` / `npx` / plugin marketplace. - No wrapper. No invented metrics. -- SIGNAL gliner-native-runtime (`research/notes.md` §97): - **Skip Archer.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#14**. Do **not** re-fold 1639 - / §96. How-to-apply / mental models / architecture / - Harbor-jevals / toolbelt — not a thin Jev skill dump. - Backend-agnostic categorization/scoring/decision class. - User-linked SIGNAL (pushed 2026-08-26). Formal methods - compose with scoring; span confidence is a SENSOR; - treating this as TypeSafe Choice/Score/Noul, collapsing - it into Fastino official, or hard-gating default 0.1 as - NER quality is soundness theater. One HIGH cluster: - **GLiNER2 native Apple path** - ([shershah1024/gliner-native-runtime](https://github.com/shershah1024/gliner-native-runtime) - unofficial Swift/Core ML GLiNER 2.5-small; entity spans - + confidence; not Choice/Score/Noul; not TypeSafe; - label descriptions as schema; on-device ANE economics; - honesty locks; Apache-2.0; **4★**; HEAD `b44f661`; - README SHA `901eb063`). Soft Noul ≠ hard safety. - default threshold 0.1 still soft. Census **not provided - this hour** (not re-derived). Archer still NOT landed. - Do not copy `swift build` / Git LFS / Hub weights. - No wrapper. No invented metrics. -- Hourly 1740 HIGH (`research/notes.md` §98): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#15**. Do **not** re-fold 1639 - / §96 / gliner-native-runtime / §97 / 1541 / §95. How-to-apply - / mental models / architecture / Harbor-jevals / toolbelt — - not a thin Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; hard-gating DGP - as safety theater, hard-gating 0.4/0.2 as “concept absent,” - treating OpenRouter/TypeSafe auto-failover as one Noul - (silent FALLBACK), or pasting kev OOD 0.76 as - Jev-equivalent is soundness theater. Three HIGH clusters: **Decision Graph Protocol - envelope** - ([numerous-com/dgp](https://github.com/numerous-com/dgp) - Decision Graph Protocol frame→assess→commit; app retains - permissions/effects; Jev-first assessor-neutral; guarded - commit / receipt/next frame; assessment batching; - numerous-com/dgp ≠ TypeSafe official; Python MIT; **0★**; - HEAD `a9cb3c4`; README SHA `655fc638`); **calibrated - meaning-grep over a live tree** - ([can1357/jegrep](https://github.com/can1357/jegrep) - jegrep calibrated path+range Nouls; no - embeddings/index/daemon; ~$0.01–0.03 typical; agent --json; - can1357/jegrep ≠ Bentlybro/jevgrep ≠ uehaj/jev-semgrep; - Rust MIT; **13★**; HEAD `a280f14`; README SHA `6f241390`); - **Archer-arch fidelity + measured calibration gap** - ([jaredpalmer/kev](https://github.com/jaredpalmer/kev) - Archer-arch fidelity; kev family OOD 0.76–0.77 vs Jev 0.86; - block-causal isolation; pointer/readout CE-trained; - /v1/systemone drop-in; replica honesty; Apache-2.0; **507★**; - do not rewrite §45; Jev-omni owns the replica fold). Soft - Noul ≠ hard safety. 0★ HIGH still got a real card. Census - **not provided this hour** except Archer tracker likes - **51** (+1); lastModified UNCHANGED 2026-09-19T18:37:18Z; - Hub `archerhume/4rcherhume` HTTP **401**. Archer still NOT - landed. Do not copy keys / `uv` / `cargo` / `.env`. No - wrapper. No invented metrics. -- Hourly 2145 HIGH (`research/notes.md` §102): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#19**. Do **not** re-fold 2041 - / §101 / 1943 / §100 / 1843 / §99 / 1740 / §98 / 1639 - / §96 / gliner-native-runtime / §97 / 1541 / §95 / - jev-align *mechanism* / §93. How-to-apply / mental - models / architecture / Harbor-jevals / toolbelt — not - a thin Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating a - clean jevq run as measured separation, treating 62 - IC-Laya tests as Laya parity, letting a Score grant Tx, - pasting “Jev48 beats Jev on phishing” from AUROC, - injecting class priors as “help”, treating v4 as a - pharmacy controller, distilling Jev as teacher of - record, pasting maze 1.00 as a general System One, or - treating jev-kit exit 0 as claim truth is soundness - theater. Nine HIGH clusters: **question-linting of Jev - questions themselves** - ([yodablocks/jevq](https://github.com/yodablocks/jevq) - PRIMARY; nine jaggedness rules, no API key, no labelled - data; static lint ≠ measured separation; - yodablocks/jevq ≠ tenbin ≠ JevLint ≠ commitjev; Python - MIT; **0★**; HEAD `40b2dd90`; README SHA `3198dde0`); - **open-weights Laya as class exemplar (binding)** - ([ChristianAlexander/laya_ex](https://github.com/ChristianAlexander/laya_ex) - Nx/Bumblebee runtime; host chooses backend; Elixir - Apache-2.0; **0★**; HEAD `99f9ce73`; README SHA - `6361c920`; mix **0.1.0**); - **on-chain/edge Laya deploy** - ([humandebri/IC-Laya](https://github.com/humandebri/IC-Laya) - parity_verified stays false; model output never grants - Tx; Rust MIT; **0★**; HEAD `055ef42f`; README SHA - `85431606`); **auditable weekend replica** - ([agilabs-ai/jev48](https://github.com/agilabs-ai/jev48) - Jev outputs never used for training; unpaired 0.577 vs - 0.727; Python MIT; **0★**; HEAD `aa697005`; README SHA - `3f667dd4`); **adversarial dual-judge / framing** - ([copyleftdev/ember](https://github.com/copyleftdev/ember) - comparative framing is the usable judgment; prior - injection crowds out evidence; TypeScript MIT; **0★**; - HEAD `c02f622b`; README SHA `6db00b56`); **Laya - specialist + Hub replica** - ([PIXELZX0/XERON](https://github.com/PIXELZX0/XERON) - training still GPU-pending; license null; **0★**; HEAD - `5e870a4d`) + - ([daliborsb/laya](https://huggingface.co/daliborsb/laya) - Hub Laya replica drop; ≠ convaiinnovations/laya); - **distillation economics** - ([MagaBitmex/jev-4b-distill-data](https://huggingface.co/datasets/MagaBitmex/jev-4b-distill-data) - gold is programmatic; teacher is closed-API clone; do - not distill Jev as teacher of record; student checkpoint - missing); **non-LLM VIN System One** - ([lewislululu/jevon](https://huggingface.co/lewislululu/jevon) - planning depth not chat; maze 1.0000 n=141 *theirs*; - AGPL; likes **3**; ≠ douglance/jevon); **source-bound - evidence** - ([WaynezProg/jev-kit](https://github.com/WaynezProg/jev-kit) - local quote mismatch needs no API; exit 0 ≠ claim - truth; MIT; **0★**; HEAD `4558554f`; README SHA - `a7f14838`; ≠ jonathanavis96/jev-kit Airlock). Soft - Noul ≠ hard safety. 0★ HIGH still - got a real card. Census **live REST pulse** (review - relock): SemIf **2000★** (+16 vs §101 **1984**); - jevlike **1002★** **flat**; TypeAR **12★** (+1 vs - §101 **11**). Tracker likes **54**; - lastModified `2026-09-20T02:59:13Z`; Hub - `archerhume/4rcherhume` HTTP **401** (not re-fetched as - a rewrite). Archer still NOT landed. Awesomejev - 561/27007 user-provided (≠ AnotiaWang/awesome-jev - **86★**). Qwen3.8-27B ≠ Archer. - `invented_signal: false`. Do not copy keys / `npm` / - `pip` / `npx` / `uv` / `mix` / `curl | sh`. No wrapper. - No invented metrics. -- Hourly 2246 HIGH (`research/notes.md` §103): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#20**. Do **not** re-fold 2145 - / §102 / 2041 / §101 / 1943 / §100 / 1843 / §99 / 1740 - / §98 / 1639 / §96 / gliner-native-runtime / §97 / 1541 - / §95 / jev-align *mechanism* / §93. How-to-apply / - mental models / architecture / Harbor-jevals / toolbelt - — not a thin Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating a - catalog row as a bake-off win, treating constructed - scenes as production logs, pasting Hev 80.00% as Jev - identity, quoting mini-jev 93.25% as family-disjoint, - putting Jev in an AR next-token loop, laundering a Noul - as a Lean step, treating parallel rank-k as a sort - proof, pasting a curated list’s von sub-15ms as an - Augustus fact, or hard-thresholding paper-radar 0.5 as - frequency is soundness theater. ChatJev-style soundness - theater is the anti-pattern. Nine HIGH clusters: - **independent System One evidence catalog** - ([reachjalil/system-one-bench](https://github.com/reachjalil/system-one-bench) - PRIMARY; 19 reviewed records; scores not one leaderboard; no external record - currently reproduced; TokenTrim no-Jev matched hybrid - 62.4%; JavaScript MIT; **0★**; HEAD `ceb17269`; README - SHA `d9e0c7b7`; ≠ mallahyari/system-one-benchmark); - **typed eval freeze** - ([SivletLabs/jev-eval](https://github.com/SivletLabs/jev-eval) - 21 tasks · 134 items · 208 questions; scenes from public - GitHub contracts, not production logs; Python MIT; - **0★**; HEAD `3f9d976f`; README SHA `df16766c`; ≠ - willkelly/jev-evaluation ≠ 4esv/jev-eval); - **option-isolated tiny replica** - ([nafisazizir/hev](https://github.com/nafisazizir/hev) - option isolation (sibling-blind); permutation-equivariant; - Hub OWNER not published; Apache-2.0; **0★**; HEAD - `79a486f9`; README SHA `b995dce3`; ≠ jaredpalmer/kev); - **frozen-LLM typed decisions** - ([yuki-oshio/mini-jev](https://github.com/yuki-oshio/mini-jev) - frozen local LLM logits, no trained decision head; - residual-head 9,222-param decreased 73/96→67/96; - confidence = 1−normalized entropy, not P(correct); - Python MIT; **0★**; HEAD `dff5b323`; README SHA - `363441b6`; ≠ r-ms/mini-jev); - **AR next-token anti-pattern** - ([erik-dunteman/ChatJev](https://github.com/erik-dunteman/ChatJev) - Jev classifier as autoregressive next-token predictor; - ChatJev-style soundness theater; license null; **1★**; - HEAD `ea33ab8d`; README SHA `c763be19`; ≠ dannote/jev ≠ - jev-gpt); - **formal compose with scoring** - ([wufuju2023-cell/jev-alpha-proof-analysis](https://github.com/wufuju2023-cell/jev-alpha-proof-analysis) - calibrated decision head × AlphaProof value head; - implementation-layer isomorphism, semantic difference; - timeout = censoring; do not launder Noul as proof; - license null; **0★**; HEAD `afd9bb6f`; README SHA - `1810d7f6`); - **parallel rank vs serial selection** - ([zzzzzec/jevsort](https://github.com/zzzzzec/jevsort) - parallel rank-prediction vs serial selection; - independent questions can conflict; HTML; license null; - **1★**; HEAD `57067b90`; README SHA `85044740`; ≠ - keltokhy/jsort); - **open-side ecosystem catalog** - ([rupeshpoojary9/awesome-open-system-one](https://github.com/rupeshpoojary9/awesome-open-system-one) - curated open System One ecosystem catalog; CC0 1.0; - SPDX NOASSERTION; **0★**; HEAD `637ee3d3`; README SHA - `0007e343`; ≠ AnotiaWang/awesome-jev); - **knowledge-work paper radar** - ([LYchoon/paper-radar-jev](https://github.com/LYchoon/paper-radar-jev) - arXiv paper radar with Jev relevance scoring; ranking ≠ - calibration / 0.5 still soft; fail-open failed evals - not marked seen; Python MIT; **0★**; HEAD `fbadf01c`; - README SHA `1cb8a9c3`; size **73**; default master). - Soft Noul ≠ hard safety. 0★ HIGH still got a real card - (jevsort **1★**). Census **live REST pulse** (review - relock): SemIf **2019★** (+19 vs §102 **2000**); - jevlike **1006★** (+4 vs §102 **1002**); TypeAR **12★** - **flat**. Tracker likes **54**; lastModified - `2026-09-20T02:59:13Z`; Hub `archerhume/4rcherhume` - HTTP **401** (not re-fetched as a rewrite). Archer - still NOT landed. Awesomejev 561/27007 user-provided - (≠ AnotiaWang/awesome-jev **87★**). Qwen3.8-27B ≠ - Archer. `invented_signal: false`. Do not copy keys / - `npm` / `pip` / `npx` / `uv` / `.env`. No wrapper. - No invented metrics. -- Hourly 2041 HIGH (`research/notes.md` §101): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#18**. Do **not** re-fold 1943 - / §100 / 1843 / §99 / 1740 / §98 / 1639 / §96 / - gliner-native-runtime / §97 / 1541 / §95 / jev-align - *mechanism* / §93. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; treating a - zero binary name gap as a fairness certificate, letting - Jev emit the verdict string, pasting PrismNLI's lead - without the contamination caveat, treating J7 pass as - safe to obey, treating a receipt as authorization, or - hard-gating confidence ≥0.95 is soundness theater. Nine - HIGH clusters: **resume-screening bias audit** - ([natemoo-re/bias-bench](https://github.com/natemoo-re/bias-bench) - PRIMARY; name×resume factorial independent Nouls; - callback determined by resume quality; mean-probability - name gaps operationally negligible; natemoo-re/bias-bench - ≠ BBQ; JavaScript; license null; **0★**; HEAD - `fe2f2535`; README SHA `a1c3e604`); **MCDA panel - code-owned verdict** - ([austindixson/planalyzer](https://github.com/austindixson/planalyzer) - Plan/PRD panel → code-owned pass|review|block; - cheerleading out of scope; Python MIT; **0★**; HEAD - `39fc161f`; README SHA `6e4d8da3`); **EU cost-aware - routing** - ([cannacre8ive/switchboard-ai](https://github.com/cannacre8ive/switchboard-ai) - cost-aware multi-model routing/escalation; decide vs do; - successful-task cost; JavaScript MIT; **1★**; HEAD - `5cae9d1c`; README SHA `872de837`; package **0.4.0**); - **frozen-protocol class bake-off** - ([elcronos/jev-vs-open-decision-models](https://github.com/elcronos/jev-vs-open-decision-models) - TypeSafe Jev vs PrismNLI vs Laya; contamination caveat; - Python; license null; **0★**; HEAD `b61e6cfc`; README - SHA `b7256888`); **VOI admission** - ([cvsgireesh/jevusher](https://github.com/cvsgireesh/jevusher) - context-window admission control; fail polarity per - lens; on small inputs lenses lose money; TypeScript MIT; - **0★**; HEAD `d830d344`; README SHA `428a4a59`); - **Leveson control plane** - ([MokiMeow/jev-fabric](https://github.com/MokiMeow/jev-fabric) - typed decision control plane; receipt ≠ authorization; - historical-v0 zero retained cases; Apache-2.0; **0★**; - HEAD `95b9a4f3`; README SHA `f485dbdd`); **scoring - economics** - ([jose-troche/live-rubric](https://github.com/jose-troche/live-rubric) - live 15-dim typed rubric re-score per pause; OpenJev/Codiv - ≠ TypeSafe hosted; ~$0.000004 desc / ~$0.000006 README; - TypeScript; license null; **0★**; HEAD `db8da8db`; - README SHA `4a0be084`); **pre-registered calibration - science** - ([willkelly/jev-evaluation](https://github.com/willkelly/jev-evaluation) - 28 predictions before data; 123,805 requests; confidence - does not track ignorance; polite injection 65% / crude - 0%; Python MIT; **0★**; HEAD `c168c093`; README SHA - `2d66ac22`; rh-guard owns injection); **class - infrastructure SDK** - ([nshkrdotcom/system_one_sdk](https://github.com/nshkrdotcom/system_one_sdk) - provider-neutral Elixir/BEAM Noul/Choice/Score SDK; - GitHub desc provider-neutral / README TypeSafe-first; - Elixir MIT; **0★**; HEAD `c2a522ee`; README SHA - `c117b4c4`; mix **0.5.0**). Soft Noul ≠ hard safety. - 0★ HIGH still got a real card (switchboard now **1★** - live REST). Census **live REST pulse**: SemIf **1984★**; - jevlike **1002★**; TypeAR **11★** flat. Tracker likes - **54** (+3 vs §100 pin **51**); lastModified **CHANGED** - 2026-09-20T02:59:13Z; Hub `archerhume/4rcherhume` HTTP - **401** (not re-fetched as a rewrite). Archer still NOT - landed. Awesomejev 561/27007 user-provided (≠ - AnotiaWang/awesome-jev 84★). - `invented_signal: false`. Do not copy keys / `npm` / - `pip` / `npx` / `uv` / `mix`. No wrapper. No invented - metrics. -- Hourly 1943 HIGH (`research/notes.md` §100): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#17**. Do **not** re-fold 1843 - / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime - / §97 / 1541 / §95 / jev-align *mechanism* / §93. - How-to-apply / mental models / architecture / - Harbor-jevals / toolbelt — not a thin Jev skill dump. - Backend-agnostic categorization/scoring/decision class. - Formal methods compose with scoring; a Noul is a SENSOR; - hard-gating `feels` 80%, pasting 17/18 as Harbor, - hard-gating 0 of 157, treating boolean @ 0.5 as a proof, - claiming 10×, treating softmax as a Noul, or treating 62 - tests as quality is soundness theater. Six HIGH clusters: - **Jev IS the if-statement** - ([southpolesteve/probably](https://github.com/southpolesteve/probably) - PRIMARY; judgments/probabilities drive branches; text - model only writes prose; interpreter owns - variables/loops/budgets/replay; otherwise maybe / - confidence gate; chaos samples after the gate; - southpolesteve/probably ≠ carldaws/hunch ≠ feelings ≠ - Kungie/gut ≠ Illusion47586/judge ≠ tidymodels/probably; - TypeScript MIT; **3★**; HEAD `6bf671a4`; README SHA - `c28570a9`); **GEPA live delta** - ([sutro-sh/jev-align](https://github.com/sutro-sh/jev-align) - 133★ / forks 10 live; build calibrated classifiers from - human feedback; HEAD `49753df9` / README SHA `363fccb7` - unchanged vs §93); **retrieve by relevance not - resemblance** - ([samdotmak/jev-recall](https://github.com/samdotmak/jev-recall) - one calibrated yes/no per memory in one request; pointer - mode 17/18 19/20 *theirs*; embedding resemblance misses - the allergy; MIT; **6★**; HEAD `d3e4acfb`; README SHA - `ea774519`); **memory leases ended by new evidence** - ([chopratejas/invalidate](https://github.com/chopratejas/invalidate) - HIGH upgrade; six Nouls then fixed rules in code; 0 of - 157 false invalidations; questions/plans/directives are - not evidence; unsure → review queue; host keeps the - store; Apache-2.0; **11★**; HEAD `d6ade601`; README SHA - `8cccad5f`); **contract-of-artifact lint rename** - ([mizchi/jev-lint](https://github.com/mizchi/jev-lint) - is mizchi/jevlint rename; name↔body / comment truth / - test-claims; no shipped rule has severity error; ~1 in 5 - findings wrong *theirs*; TypeScript MIT; **13★**; HEAD - `62d73f8e`; README SHA `4c0e37cd`); **JSON Schema - question compiler** - ([Kiln-AI/jev_jsonschema](https://github.com/Kiln-AI/jev_jsonschema) - HIGH upgrade; JSON Schema → typed JSON via Jev; - noul_threshold 0.5 decoder not a proof; - IncompatibleSchemaError lists every bad property; MIT; - **5★**; HEAD `fccea8c2` unchanged); **local System One - economics** - ([mizorewww/laya-coreml](https://github.com/mizorewww/laya-coreml) - on-device Laya CoreML ANE; ~5 ms P50 short decisions; - 189/189 FP16 checkpoint parity; 10× not achieved; - Apache-2.0; **0★**; HEAD `47f4baf0`; README SHA - `2068c661`) + - ([Micha0827/snapjudge](https://github.com/Micha0827/snapjudge) - softmax over allowed tokens ≠ Noul; question-first - cache; Python MIT; **3★**; HEAD `2df5ce27`; README SHA - `64a91458`) + - ([direwolfiy/JevPi](https://github.com/direwolfiy/JevPi) - Jev-first Pi agent loop; slow-LLM fallback; explicit - action menu / CandidateSource unimplemented; 62 tests - wiring not quality; license null; **0★**; HEAD - `980f8895`; README SHA `88ec1a55`). Soft Noul ≠ hard - safety. 0★ HIGH still got a real card. Census **live - REST pulse**: SemIf **1954★**; jevlike **989★**; TypeAR - **11★** flat. Tracker likes **51** flat; lastModified - UNCHANGED 2026-09-19T18:37:18Z; Hub - `archerhume/4rcherhume` HTTP **401** (not re-fetched as - a rewrite). Archer still NOT landed. Awesomejev - 561/27007 user-provided (≠ AnotiaWang/awesome-jev 83★). - `invented_signal: false`. Do not copy keys / `bun` / - `pip` / `npx` / `uv` / HF download. No wrapper. No - invented metrics. -- Hourly 1843 HIGH (`research/notes.md` §99): **Skip - Archer rewrite.** Docs-only on a **fresh PR off main**. - Never reopen merged #7–**#16**. Do **not** re-fold 1740 - / §98 / 1639 / §96 / gliner-native-runtime / §97 / - 1541 / §95. How-to-apply / mental models / - architecture / Harbor-jevals / toolbelt — not a thin - Jev skill dump. Backend-agnostic - categorization/scoring/decision class. Formal methods - compose with scoring; a Noul is a SENSOR; hard-gating - cost-derived 0.038 as a proof, merging von Needle - 52.6% with n=78 93%, “guaranteeing” calibration, or - pasting “Jev wins guardrailing” is soundness theater. - Five HIGH clusters: **cost-derived YES/NO/UNSURE - control flow** - ([Kungie/gut](https://github.com/Kungie/gut) - PRIMARY; cost-sensitive decision theory × System One - probabilities → control flow; thresholds derived from - costs not hard-coded; YES / NO / UNSURE from - cost_false_yes / cost_false_no / cost_human; - auto-batching same-object questions; Kungie/gut ≠ - tpellet/hunch ≠ carldaws/hunch; GitHub Apache-2.0 / - LICENSE MIT / pyproject Apache-2.0 *theirs*; **0★**; - HEAD `cb56c875`; README SHA `630474f6`; pre-alpha); - **typed-callback twin** - ([Illusion47586/judge](https://github.com/Illusion47586/judge) - judgment vs generation; deterministic execution after - probabilistic judgment; exactly one app-owned - callback; explicit uncertain branch; - Illusion47586/judge ≠ lexingtonhibiki/judgekit ≠ - Ascurse/typed-judge-kit; TypeScript MIT; **0★**; HEAD - `e69f65a1`; README SHA `08554c6f`; `@brkn-labs/judge` - 0.1.0); **variable-N option scoring as the trainable - object** - ([zwliJay/jev-forge](https://github.com/zwliJay/jev-forge) - dynamic candidate bags not fixed label sets; - zwliJay/jev-forge ≠ NanoJev; class-architecture note, - not a sixth species; GitHub NOASSERTION / LICENSE MIT; - **1★**; HEAD `eb3e4a2d`; README SHA `359f3f57`; do not - clone Hub weights); **open NAR replica economics** - ([wfzyx/von](https://github.com/wfzyx/von) late-catch - HIGH of the same repo; NAR local drop-in; open replica - economics / latency vs closed Jev; competing NAR - claims / replica honesty; Apache-2.0; **43★**; HEAD - `b9e42b26`; README SHA `574aa628`; do not merge Needle - 52.6% with n=78 93.0%); **typed vs chat judges on - guardrailing** - ([ishaannk/llm-vs-jev](https://github.com/ishaannk/llm-vs-jev) - cross-note only; nothing wins outright; can be argued - out of guarding; deeper integrity fold is rh-guard; - Apache-2.0; **0★**; HEAD `182e0864`; README SHA - `d9ebd40f`). gut/judge are **control-flow / - decision-theory overlays, not new class-table - species**. Soft Noul ≠ hard safety. 0★ HIGH still got - a real card. Census **live REST quoted**: SemIf - **1936★**; jevlike **983★** (watch claimed 984); - TypeAR **11★** flat. Tracker likes **51** flat; - lastModified UNCHANGED 2026-09-19T18:37:18Z; Hub - `archerhume/4rcherhume` HTTP **401**. Archer still NOT - landed. Awesomejev 561/27007 user-provided (≠ - AnotiaWang/awesome-jev 83★). X MCP `since_id` held; - pages_archived 0; no invented tweets. - `invented_signal: false`. Do not copy keys / `pip` / - `npm` / `uv` / Hub download. No wrapper. No invented - metrics. - Effect-oriented loops (`notes.md` §28, `mappings.md` §19): Ward's ZIO client keeps Jev as the outer Choice and the handler as the effect. Not Effect.ts. GLiNER author: GLiNER2 "like jev" is GLiGuard diff --git a/CITATION.cff b/CITATION.cff new file mode 100644 index 00000000..5de8e29c --- /dev/null +++ b/CITATION.cff @@ -0,0 +1,28 @@ +cff-version: 1.2.0 +title: Augustus +message: If you use this skill, please cite it using these metadata. +type: software +authors: + - name: Augustus contributors + website: https://github.com/24601/Augustus + - family-names: Mustafa + given-names: Basit + alias: "24601" +repository-code: https://github.com/24601/Augustus +url: https://24601.github.io/Augustus/ +abstract: >- + Design-judgment skill for placing typed probabilistic judgment + (the Jev-class of System One / decision models) using mathematical, + logical, and algorithmic mental models. TypeSafe Jev is the + documented exemplar, not the monopoly. Formal methods are one pillar. + Not a TypeSafe product. +keywords: + - system-one + - decision-judgment + - jev-class + - zero-shot-classification + - formal-methods + - agentic-ai +license: MIT +version: 0.4.0 +date-released: "2026-09-20" diff --git a/CODE_OF_CONDUCT.md b/CODE_OF_CONDUCT.md new file mode 100644 index 00000000..e930c243 --- /dev/null +++ b/CODE_OF_CONDUCT.md @@ -0,0 +1,12 @@ +# Code of conduct + +Be respectful. Do not harass anyone. Argue about placements and evidence, +not people. + +This is a small MIT project. Maintainers may reject contributions that +are hostile, that invent metrics, or that treat a soft score as a hard +safety gate. + +Report conduct problems the same way as other repo issues — or privately +via GitHub Security Advisories if the report itself should stay off the +public tracker. There is no separate conduct email. diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md new file mode 100644 index 00000000..9212df5b --- /dev/null +++ b/CONTRIBUTING.md @@ -0,0 +1,56 @@ +# Contributing + +Docs and skill PRs are welcome. This is a placement skill, not a TypeSafe +product and not a vendor how-to. + +## What to send + +- Skill / reference-card clarifications that keep Jev as **exemplar, not + monopoly** +- Measurement-honesty fixes (ranking ≠ calibration; Score is 0..n−1; + Noul has no confidence field; soft Noul ≠ hard gate) +- Composition notes (fail-open vs fail-closed per action; prune ≠ deny) +- Formal-methods honesty (a Noul is a SENSOR, not a proof) + +Do not invent metrics. Do not paste `pipeline()` / `pip` / `npm` install +recipes that read as endorsements. Quote *theirs* and label vendor +figures. + +## Adversarial review + +Read the change against the skill's own non-negotiables before opening +the PR. The usual failure modes are soundness theater: hard-gating a +soft Noul, treating ECE as an edge, treating 0.85 / minProbability as +Harbor, or laundering a score as a proof. + +Run what you can locally: + +```bash +python3 .agents/skills/augustus/scripts/evaluate_decisions.py --self-test +``` + +The Pages workflow must stay green (`docs/index.md` still contains the +gate strings the check greps for). + +## Fold uniqueness + +Hourly research folds carry uniqueness locks so the same cluster is not +re-opened as "new." Before folding: + +- Read `research/notes.md` and the uniqueness fragments in + `.agents/skills/augustus/SKILL.md` +- Do not re-fold an already-landed section as a new beat +- Do not reopen or amend an in-flight fold PR (including open #31) +- Pre-0.4.0 uniqueness dump: `research/changelog-hourly.md` (archive, + not release notes) + +## Secrets + +No API keys, tokens, or `.env` files in the tree. Push protection and +secret scanning are on. See `SECURITY.md`. + +## Branch protection + +Default-branch protection, required checks, and org settings are +parent-owned. This file does not change them. Prefer a PR off latest +`main`. Do not force-push shared branches. diff --git a/README.md b/README.md index 3a09bd9d..172c795e 100644 --- a/README.md +++ b/README.md @@ -1,10 +1,16 @@ # Augustus +Place typed probabilistic judgment — Jev-class System One / decision +models — using classical mental models. Jev is the exemplar, not the monopoly. + [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE) -[![Version](https://img.shields.io/github/v/release/24601/Augustus)](https://github.com/24601/Augustus/releases) +[![Release](https://img.shields.io/github/v/release/24601/Augustus)](https://github.com/24601/Augustus/releases) +[![Pages](https://img.shields.io/badge/docs-24601.github.io-blue.svg)](https://24601.github.io/Augustus/) [![Claude Code](https://img.shields.io/badge/Claude_Code-marketplace-purple.svg)](.claude-plugin/marketplace.json) [![Skills.sh](https://img.shields.io/badge/skills.sh-compatible-green.svg)](https://www.skills.sh/) +**Homepage:** [24601.github.io/Augustus](https://24601.github.io/Augustus/) + Agent skill for placing TypeSafe Jev Choice/Score/Noul with classical decision methods, composition algebra, and a validation gate. @@ -18,12 +24,13 @@ mathematical, logical, and algorithmic mental models. It applies across Exact work stays in code or policy; the model owns narrow judgment; never launder a Noul as a proof. -> Companion, not replacement, to the official +> **Not a TypeSafe product.** Companion, not replacement, to the official > [`typesafe-ai` skill](https://github.com/typesafe-ai/skills). That skill > owns Jev integration contracts; Augustus owns the **design judgment**: > which *pillar*, *family*, and classical method map, what the objective > implies for fail-open vs fail-closed, and what experiment would prove a -> design wrong. Not a TypeSafe-only how-to. +> design wrong. Not a TypeSafe-only how-to. Integrity / reward-hack +> companion: [`rh-guard`](https://github.com/24601/rh-guard). ## The skill @@ -400,6 +407,14 @@ claude plugin install augustus@augustus npx skills add 24601/Augustus --skill augustus ``` +**Copy the skill path** (Cursor / Amp / any agent that reads repo-local +skills): + +```bash +git clone https://github.com/24601/Augustus.git +# skill lives at .agents/skills/augustus/ +``` + **ChatGPT**: skills are not a native ChatGPT primitive — paste `.agents/skills/augustus/SKILL.md` plus the `references/` files into a GPT's instructions or a Project's knowledge and it will follow the protocol. @@ -411,21 +426,23 @@ GPT's instructions or a Project's knowledge and it will follow the protocol. `jev` `typesafe` `typesafe-ai` `system-one` `system-one-models` `structured-output` `calibrated-confidence` `ai-agents` `agent-skills` `decision-systems` `reranking` `beam-search` `claude-code` `python` `llm` -`decision-theory` `semantic-search` `agent-workflows` `mixed-architecture` +`decision-theory` `decision-making` `semantic-search` `agent-workflows` +`mixed-architecture` `agentic-ai` `zero-shot-classification` `tool-routing` `skill-routing` `semantic-lint` `classification` `gliclass` `listwise-ranking` `vision-scoring` `open-weights` `formal-methods` -`model-checking` `deterministic-simulation` `decision-theory` +`model-checking` `deterministic-simulation` `value-of-information` `signal-detection` `mcda` `calibration` `alloy` `apalache` `pufferlib` `stamp-stpa` ## Versioning See [CHANGELOG.md](CHANGELOG.md) and -[releases](https://github.com/24601/Augustus/releases). Current: **0.3.0**, +[releases](https://github.com/24601/Augustus/releases). Current: **0.4.0**, written against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) -(`65a39f3`). Re-read live TypeSafe docs before treating that pin as current -API behavior. +(`65a39f3`; live HEAD still this commit). Re-read live TypeSafe docs +before treating that pin as current API behavior. ## License -MIT — see [LICENSE](LICENSE). +MIT — see [LICENSE](LICENSE). Security reports: [SECURITY.md](SECURITY.md). +Contributions: [CONTRIBUTING.md](CONTRIBUTING.md). diff --git a/SECURITY.md b/SECURITY.md new file mode 100644 index 00000000..e61a0739 --- /dev/null +++ b/SECURITY.md @@ -0,0 +1,41 @@ +# Security + +## Supported versions + +| Version | Supported | +| ------- | --------- | +| 0.4.x | Yes | +| 0.3.x | No | +| < 0.3 | No | + +This repo is a design-judgment skill plus offline scripts. There is no +hosted API and no runtime that accepts untrusted input by default. + +## Report a vulnerability + +Use GitHub **Security Advisories** / private vulnerability reporting on +this repository: + +https://github.com/24601/Augustus/security/advisories/new + +Do not open a public issue for a security report. There is no separate +security email. + +Please include what is affected (skill text, evaluator, workflows, Pages) +and a way to reproduce. We will acknowledge and ship a fix on the +supported line when the report is valid. + +## Already enabled (do not disable) + +- Dependabot security updates +- Secret scanning +- Push protection for secrets + +## Secrets + +Do not commit API keys, tokens, or `.env` files. `.gitignore` covers the +common names. Research folds quote public READMEs; they are not install +recipes and must not paste live credentials. + +`evaluate_decisions.py --self-test` is local scoring math. It does not +call a model and does not need a key. diff --git a/docs/_config.yml b/docs/_config.yml index b9c59ce2..3253553a 100644 --- a/docs/_config.yml +++ b/docs/_config.yml @@ -1,7 +1,8 @@ title: Augustus description: >- - Agent skill for placing TypeSafe Jev Choice/Score/Noul with classical - decision methods, composition algebra, and a validation gate. + System One decision-judgment skill for the Jev-class of typed + probabilistic models. Place TypeSafe Jev Choice/Score/Noul — Jev is + the exemplar, not the monopoly — beside code, policy, and proof. url: https://24601.github.io baseurl: /Augustus theme: jekyll-theme-cayman diff --git a/docs/index.md b/docs/index.md index 6999fafd..f835b315 100644 --- a/docs/index.md +++ b/docs/index.md @@ -1,11 +1,18 @@ --- layout: default -title: Home +title: Augustus — System One decision judgment permalink: / --- -Agent skill for placing TypeSafe Jev Choice/Score/Noul with classical -decision methods, composition algebra, and a validation gate. +**Last updated:** 2026-09-20 (v0.4.0) + +Agent skill for placing TypeSafe Jev Choice/Score/Noul — and the wider +Jev-class of System One / decision models — with classical decision +methods, composition algebra, and a validation gate. + +v0.4.0: Jev is the exemplar, not the monopoly (Laya, kev, OpenJev, +TypeAR, GLiNER, encoder zero-shot). Ranking is not calibration. A soft +Noul is not a hard gate. Formal methods stay a pillar. - [Install the skill](https://github.com/24601/Augustus#install) - [Ecosystem](ecosystem.md) @@ -19,4 +26,5 @@ Companion to the official [`typesafe-ai` skill](https://github.com/typesafe-ai/skills) (Jev contracts). Augustus owns **where** judgment belongs: pillar, family, fail polarity, and the experiment that could prove a design -wrong. +wrong. Not a TypeSafe product. Integrity / reward-hack companion: +[`rh-guard`](https://github.com/24601/rh-guard). diff --git a/docs/release-notes-v0.4.0.md b/docs/release-notes-v0.4.0.md new file mode 100644 index 00000000..25308300 --- /dev/null +++ b/docs/release-notes-v0.4.0.md @@ -0,0 +1,40 @@ +Design judgment across domains. Augustus places typed probabilistic +judgment (Jev is the exemplar, not the monopoly) using math, logic, and +algorithmic mental models. The job is not limited to software +engineering. Formal methods are one pillar. Exact work stays in code or +policy; the model owns narrow judgment; a soft Noul is not a proof. + +Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) +(`65a39f3`, 2026-09-12). Live HEAD of that repo is still this commit. + +### Added + +- **Class breadth.** Laya, kev, OpenJev, TypeAR, GLiNER species, Decision + Graph Protocol, encoder zero-shot classifiers. Jev vs GPT-5.6 bakeoffs + are a category error (Merve: BERTForXYZ → DeBERTa → ModernBERT). +- **Measurement honesty.** Harbor / jevals practice: ranking ≠ calibration. + Score is 0..n−1. Noul has no confidence field. ECE ≠ an edge. 0.85 / + minProbability theater. Soft Noul ≠ hard gate. VERIFY needs + discriminating evidence. +- **Composition.** DGP frame→assess→commit; meaning-grep; gut + cost-of-error; fail-open vs fail-closed; compaction 0.5 is not safety; + prune ≠ deny. +- Community-health stubs (SECURITY, CONTRIBUTING, CODE_OF_CONDUCT) and + an OpenSSF Scorecard workflow so a score can appear. No invented + Scorecard number. + +Formal methods pillar unchanged in spirit. Homepage: +https://24601.github.io/Augustus/ + +### Install + +```bash +claude plugin marketplace add 24601/Augustus +claude plugin install augustus@augustus +``` + +```bash +npx skills add 24601/Augustus --skill augustus +``` + +Full notes: [CHANGELOG.md](https://github.com/24601/Augustus/blob/main/CHANGELOG.md) diff --git a/research/README.md b/research/README.md index f829f1d4..8d6b9ed8 100644 --- a/research/README.md +++ b/research/README.md @@ -6,6 +6,8 @@ refreshes diff against a known baseline instead of re-discovering the world. - `sources.json` — every source pulled, with type + retrieval date + note. - `notes.md` — distilled findings (contracts, recipes, ecosystem, gaps). - `refresh-log.md` — dated log of each refresh pass and what changed. +- `changelog-hourly.md` — pre-0.4.0 uniqueness-lock dump (not release + notes; see root `CHANGELOG.md`). - `archive/hourly/YYYY-MM-DDTHH/` — raw scan dumps for that UTC hour (X theme digest + `topic:jev` JSON when a live scan lands). - `archive/curriculum/` — attached research briefs folded into the skill diff --git a/research/changelog-hourly.md b/research/changelog-hourly.md new file mode 100644 index 00000000..bca5ad57 --- /dev/null +++ b/research/changelog-hourly.md @@ -0,0 +1,3556 @@ +# Hourly uniqueness dump (pre-0.4.0) + +This is the pre-0.4.0 `CHANGELOG.md` after hourly folds (#2–#30 / notes +§44–§112) stuffed uniqueness locks into Keep-a-Changelog sections. It is +**not** release notes. + +- Release-facing notes: [`CHANGELOG.md`](../CHANGELOG.md) +- Canonical folds: [`notes.md`](notes.md) +- Do not treat numbers here as new measurements. Quote *theirs*. +- Soft Noul ≠ hard gate. Ranking ≠ calibration. + +--- + +# Changelog + +All notable changes to Augustus are documented here. The format follows +[Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and versioning +follows [SemVer](https://semver.org/spec/v2.0.0.html). + +Each release notes the [`typesafe-ai/skills`](https://github.com/typesafe-ai/skills) +revision it was written against. That skill owns integration contracts; +Augustus owns design judgment. Re-read live TypeSafe docs before treating a +pin as current API behavior. + +## [0.3.0] - 2026-09-18 + +Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) +(`65a39f3`, 2026-09-12). + +### Added + +- User-provided HIGH Merve Noyan ZS classifier lineage + (`research/notes.md` §112): **Skip Archer rewrite.** + Docs-only on a **fresh PR off main**. + **HARD RULE:** do not reopen or amend PR #23 or #24 or + #25 or #26 or #27 or #28 or #29 (merged #29 owns + `notes.md` §111 / items 248–267 / batch #94 — leave it + alone). Institutional HF voice (@mervenoyann). Quote + *theirs*. Do not invent accuracy numbers. Jev vs + GPT-5.6 bakeoffs are a category error. encoder / ZS + classifiers (BERTForXYZ → DeBERTa → ModernBERT). many + problems solved with LLMs could have been solved with + them, it was a skill issue. opt for DeBERTa and + ModernBERT ones. multimodal image<>text ZS as + perception front-end. softmax/ZS scores still ≠ + calibrated Noul. soft scores ≠ hard gates. Live X this + pass: likes 421 / 189; impressions 35498 / 9613. + Hub likes 139 / 72 cited. X MCP used. + `invented_signal: false`. Composition items 268–272 / + batch #95. Uniqueness lock: people who compare Jev against GPT-5.6 has never fine-tuned BERTForXYZ for living and it shows; zero shot classifiers; scale them as much as decoder only models; many problems solved with LLMs could have been solved with them, it was a skill issue; opt for DeBERTa and ModernBERT ones; BERTForXYZ → DeBERTa → ModernBERT; Jev vs GPT-5.6 bakeoffs are a category error; encoder / ZS classifiers; institutional HF voice; quote *theirs*; do not invent accuracy numbers; softmax/ZS scores still ≠ calibrated Noul; soft scores ≠ hard gates; @mervenoyann; likes 421 / 189; impressions 35498 / 9613; multimodal image<>text ZS as perception front-end; hf:MoritzLaurer/deberta-v3-large-zeroshot-v2.0 likes 139; hf:MoritzLaurer/ModernBERT-large-zeroshot-v2.0 likes 72; Bart, bert, deberta, modernbert, these are all LLMs; Maziyar quoted; Jev is exemplar not the mandate; do not reopen or amend PR #23/#24/#25/#26/#27/#28/#29. + +- Hourly 0646 HIGH (`research/notes.md` §111): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + **HARD RULE:** do not reopen or amend PR #23 or #24 or #25 + or #26 or #27 or #28 (do not reopen or amend PR #23/#24/#25/#26/#27/#28. + merged #23 owns `notes.md` §105 / items 149–160 / + batch #88; merged #24 owns `notes.md` §106 / items 161–177 / + batch #89; merged #25 owns `notes.md` §107 / items 178–185 / + batch #90; merged #26 owns `notes.md` §108 / items 186–201 / + batch #91; merged #27 owns `notes.md` §109 / items 202–225 / + batch #92; merged #28 owns `notes.md` §110 / items 226–247 / + batch #93 — leave them alone). Never reopen merged + #7–**#28**. Do **not** re-fold 0541 / §110 / 0439 / §109 / 0345 / §108 / 0243 / §107 / + 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / §103 / + 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 / §99 / + 1740 / §98 / 1639 / §96 / gliner-native-runtime / §97 / + 1541 / §95 / jev-align *mechanism* / §93 / + jev-orderby-bench *six-gates* / §60 / JevBench v1.2 + *board* / §78 / openJev-verdict *claim-audit* / §71 / + jev-judge-bench SLA-150 *contract* / §71 / + yuki-oshio/mini-jev *93.25%* / §103. How-to-apply / mental + models / architecture / Harbor-jevals / toolbelt — not a + thin Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating + ECE as alpha, 0.5 compaction as safety, 0.8 nlgrep as + 80% correctness, contract_passed as truth, Wilson 0.85 + as a proof, 90.53% BBQ as calibrated, softmax over + A/B/C as a Noul, 0.740 as TypeSafe vs Laya, 97.0% as + Harbor, or intervals-including-zero as equivalence is + soundness theater. Calibration is not alpha and ranking ≠ + calibration theater are the anti-patterns. Twenty HIGH + clusters / three themes: + **Measurement / benches / calibration PRIMARY** + ([alakise/calibration-is-not-alpha](https://github.com/alakise/calibration-is-not-alpha) + Calibration is not alpha; NO CURRENT ALPHA CANDIDATE; + ΔR² approximately +0.00084; Brier 0.2131387; ECE 0.0421875; + Adding Jev probability to deterministic volatility improved Brier by only 1.4058e-05; + Python MIT LICENSE SHA `cbfc2daf`; **0★**; HEAD `064b75f5`; README SHA `874ea57b`; size **268**); + ([OrMizL/jev-compaction-bench](https://github.com/OrMizL/jev-compaction-bench) + default 0.5 keeps zero non pinned; + keepResult median 0.14 to 0.17; keepCall median 0.28 to 0.35; + usable range is about 0.10 to 0.25; 7.8% to 57.9%; + judges results it never sees; task-finish eval not built yet; + $0.002 per compaction; + JavaScript MIT LICENSE SHA `3f4e7f2b`; **0★**; HEAD `92fd33e6`; README SHA `c7cd8b6e`; size **35**); + ([slavadubrov/sgr-judge-bench](https://github.com/slavadubrov/sgr-judge-bench) + slavadubrov/sgr-judge-bench ≠ slavadubrov/jev-judge-bench; + Jev 108/120 $0.083 0.34 s; Luna SGR 114/120; + paired Jev accuracy-difference intervals include zero; + not evidence of equivalence; + GLM SGR 26/120 93 format failures; Terra-planned Jev hybrid 55/120; + Python; SPDX NOASSERTION LICENSE SHA `690cf6b9`; **0★**; HEAD `5e142707`; README SHA `8026cafa`; size **0** WITH CONTENTS); + ([elyashium/atlas-replay-lab](https://github.com/elyashium/atlas-replay-lab) + rule-based by default, optionally Jev-backed; empty README; + missing key cannot break the experience; + JavaScript; GitHub license null; **0★**; HEAD `9856ab9b`; size **197**); + ([siren2345/jev-single-decode](https://github.com/siren2345/jev-single-decode) + prefill plus exactly one decode; softmax over A/B/C ≠ Noul; + BBQ 9,053/10,000 (90.53%); ECE 0.0890; Mean confidence 0.9943; + overconfident; score and noul not implemented; + Python MIT LICENSE SHA `3551844d`; **0★**; HEAD `65df86a3`; README SHA `51817d8c`; size **432**); + **datasets / Spaces** + (hfdataset:ctaxnagomi/DGUI_HYPERMEM-JEV + DGUI 12 rows (was 6); sha `ab3d3529`); + (hfdataset:ctaxnagomi/INSTRUCT_JEV + INSTRUCT 119 rows likes 2; sha `b0a09278`); + (hfspace:pngwn/open-jev + encode the state once, decide everything in parallel; + 0.740 accuracy against a 0.508 majority; ECE 0.047; + fine-tune's advantage ends where its 384-token training data does; + likes **25**; sha `d41dc3cd`); + (hfspace:jasonkneen/open-jev + jasonkneen/open-jev ≠ pngwn/open-jev; same sha d41dc3cd; likes **0**); + (hfspace:IkerMoel/open-alternative-jev densify sha `19b104c6`); + (hfspace:mobarmg/jev-schema-scorer densify sha `14d35d7e`; peaked ranking ≠ calibration); + (hfspace:reachjalil/jevlogs-triage-explorer + Space does not call Jev; recomputes routing from saved probabilities; + sha `dcb785ed`); + **applied / skills / economics** + ([IslamBaraka90/jev-typesafe-real-financial-use-cases](https://github.com/IslamBaraka90/jev-typesafe-real-financial-use-cases) + fifty recorded demos; recorded ≠ alpha; + JavaScript MIT LICENSE SHA `caaa17a8`; **0★**; HEAD `9d2eb48c`; README SHA `5329b92e`; size **6168**); + ([Milo318/mailordinal](https://github.com/Milo318/mailordinal) + 200-case Jev 97.0% / 100.0% / 95.0% / MAE 9.22; + synthetic repository benchmark; + TypeScript MIT LICENSE SHA `66faa16d`; **0★**; HEAD `19ea819d`; README SHA `a8df45fe`; size **109**); + ([Pleo2/awesome-jev-agent-skills](https://github.com/Pleo2/awesome-jev-agent-skills) + Jev evaluations are advisory; catalog ≠ endorsement; + Python MIT LICENSE SHA `25d9aa30`; **0★**; HEAD `42e3d179`; README SHA `efbe7ccf`; size **18**); + ([YehuiTang0316/jev-nlgrep](https://github.com/YehuiTang0316/jev-nlgrep) + YehuiTang0316/jev-nlgrep ≠ Bentlybro/jevgrep ≠ can1357/jegrep ≠ uehaj/jev-semgrep; + default threshold 0.8 still soft; 40-line windows cannot prove whole function; + TypeScript MIT LICENSE SHA `17cdbc7e`; **1★**; HEAD `ceec0d92`; README SHA `49e98f18`; size **6734**); + ([dangquan1402/jev-extract](https://github.com/dangquan1402/jev-extract) + token-native sequential start/end Choice; Gemini/Haiku stubs not configured yet; + Python MIT LICENSE SHA `d66f9c7a`; **0★**; HEAD `20c2f19f`; README SHA `9561e53d`; size **105**); + ([jyje/pilot-typesafeai-jev](https://github.com/jyje/pilot-typesafeai-jev) + handful of hand-written examples, not a benchmark; + Jev judged exactly what it was given; + Python MIT LICENSE SHA `08b2a315`; **0★**; HEAD `cabd4778`; README SHA `8d472372`; size **245**); + ([laguagu/jev-skills](https://github.com/laguagu/jev-skills) + laguagu/jev-skills ≠ laguagu/jev-evidence-lab ≠ Pleo2/awesome-jev-agent-skills; + MIT LICENSE SHA `e5ede121`; **0★**; HEAD `871ec586`; README SHA `a315944d`; size **72**); + ([lorensation/llm-cost-optimizer-jev](https://github.com/lorensation/llm-cost-optimizer-jev) + contract_passed is not a claim of guaranteed factual truth; + Wilson lower bound 0.85 floor; fixture mode no savings claim; + Python Apache-2.0 LICENSE SHA `261eeb9e`; **0★**; HEAD `db200b7e`; README SHA `6f2a595a`; size **227**). + SemIf 2207★ (+21 vs §110 2186); + jevlike 1043★ (+5 vs 1038); + TypeAR 15★ (+1 vs 14); + AnotiaWang 97★ (+1 vs 96); + yibie/awesome-jev 506★ (+16 vs 490); + Laya likes 822 (was 802); + tracker likes 64 flat, lastModified UNCHANGED; + Laya present; Blackwood ABSENT; Archer still promised_not_landed. + 0★ HIGH still got a real card. Soft Noul ≠ hard safety. + `invented_signal: false`. + Hourly 0646 uniqueness lock: Calibration is not alpha; NO CURRENT ALPHA CANDIDATE; ΔR² approximately +0.00084; Brier 0.2131387; ECE 0.0421875; Adding Jev probability to deterministic volatility improved Brier by only 1.4058e-05; default 0.5 keeps zero non pinned; keepResult median 0.14 to 0.17; keepCall median 0.28 to 0.35; usable range is about 0.10 to 0.25; 7.8% to 57.9%; judges results it never sees; task-finish eval not built yet; $0.002 per compaction; slavadubrov/sgr-judge-bench ≠ slavadubrov/jev-judge-bench; Jev 108/120 $0.083 0.34 s; Luna SGR 114/120; paired Jev accuracy-difference intervals include zero; not evidence of equivalence; GLM SGR 26/120 93 format failures; Terra-planned Jev hybrid 55/120; rule-based by default, optionally Jev-backed; empty README; missing key cannot break the experience; prefill plus exactly one decode; softmax over A/B/C ≠ Noul; BBQ 9,053/10,000 (90.53%); ECE 0.0890; Mean confidence 0.9943; overconfident; score and noul not implemented; DGUI 12 rows (was 6); INSTRUCT 119 rows likes 2; encode the state once, decide everything in parallel; 0.740 accuracy against a 0.508 majority; ECE 0.047; fine-tune's advantage ends where its 384-token training data does; jasonkneen/open-jev ≠ pngwn/open-jev; same sha d41dc3cd; Space does not call Jev; recomputes routing from saved probabilities; 200-case Jev 97.0% / 100.0% / 95.0% / MAE 9.22; synthetic repository benchmark; Jev evaluations are advisory; YehuiTang0316/jev-nlgrep ≠ Bentlybro/jevgrep ≠ can1357/jegrep ≠ uehaj/jev-semgrep; default threshold 0.8 still soft; 40-line windows cannot prove whole function; token-native sequential start/end Choice; Gemini/Haiku stubs not configured yet; handful of hand-written examples, not a benchmark; Jev judged exactly what it was given; laguagu/jev-skills ≠ laguagu/jev-evidence-lab ≠ Pleo2/awesome-jev-agent-skills; contract_passed is not a claim of guaranteed factual truth; Wilson lower bound 0.85 floor; fixture mode no savings claim; SemIf 2207★ (+21 vs §110 2186); jevlike 1043★ (+5 vs 1038); TypeAR 15★ (+1 vs 14); AnotiaWang 97★ (+1 vs 96); yibie/awesome-jev 506★ (+16 vs 490); Laya likes 822 (was 802); tracker likes 64 flat, lastModified UNCHANGED; do not reopen or amend PR #23/#24/#25/#26/#27/#28. + +- Hourly 0541 HIGH (`research/notes.md` §110): **Skip + Archer rewrite.** Live-REST relock after adversarial + review (HEAD/README/size/stars/likes; Brier/Independent + design claims unchanged). Docs-only on a **fresh PR off main**. + **HARD RULE:** do not reopen or amend PR #23 or #24 or #25 + or #26 or #27 (do not reopen or amend PR #23/#24/#25/#26/#27. + merged #23 owns `notes.md` §105 / items 149–160 / + batch #88; merged #24 owns `notes.md` §106 / items 161–177 / + batch #89; merged #25 owns `notes.md` §107 / items 178–185 / + batch #90; merged #26 owns `notes.md` §108 / items 186–201 / + batch #91; merged #27 owns `notes.md` §109 / items 202–225 / + batch #92 — leave them alone). Never reopen merged + #7–**#27**. Do **not** re-fold 0439 / §109 / 0345 / §108 / 0243 / §107 / + 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / §103 / + 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 / §99 / + 1740 / §98 / 1639 / §96 / gliner-native-runtime / §97 / + 1541 / §95 / jev-align *mechanism* / §93 / + jev-orderby-bench *six-gates* / §60 / JevBench v1.2 + *board* / §78 / openJev-verdict *claim-audit* / §71 / + yuki-oshio/mini-jev *93.25%* / §103. How-to-apply / mental + models / architecture / Harbor-jevals / toolbelt — not a + thin Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating + Blackwood tracker-absent as landed, 0.9 as one number, + 8/8 as conversion lift, 200-row as ranking, 0.85 as 85%, + Hub Meanblock as a new species vs leesk212, circulating + $0.0004 as measured, Score as 0–1, Noul.confidence as + existing, json_schema gap as typed-model win, 64× Space + as Harbor, awesome listed counts as eval, WANLI 0.741 as + beating openjev, ECE 0.021 as a hard gate, Independent + `{cat,dog}` as Choice, softmax over letter slots as a + Noul, minProbability 0.85 as Harbor, or :max as a + mandatory search size is soundness theater. treating + 0.85 as 85% / minProbability hard-gate as Harbor, 0.9 + as one number, and 8/8 as conversion lift are the + anti-patterns. Twenty-two HIGH clusters / three themes: + **Open-weight / RLCD / Blackwood watch** + (hf:BlackwoodAI/blackwood-rlcd + Blackwood tracker ABSENT; likes 2 gated manual; + sha `3b9e29df`); + (hf:anthonym21/qwen3-0.6b-rlcd-decision + r = c - p_a; ECE 0.021; acc 0.807 vs warmup 0.746; + calibration beyond ~500 tokens unmeasured; + apache-2.0; likes **2**; sha `b327ec5e`); + (hf:larkooo/gemma-e2b-rlcd + Independent primitive; + 11.57s vs 54.10s · 4.67× · 120/128 *theirs*; + default path is pretrained Gemma probs not trained RLCD head; + apache-2.0; likes **1**; sha `e099c730`); + (hf:Meanblock/JEV-CPU + GH Meanblock 404; lock leesk212/JEV-CPU; + softmax over letter slots ≠ Noul; + mit; likes **7**; sha `759fa606`); + (hf:impacte/mimir-lfm-openjev + WANLI 0.741 vs openjev v2 0.77 *theirs*; + 3-way NLI ≠ Noul; likes **0**; sha `5f9173bb`); + (hf:shreyanbr/system-one-distilled | gold | zeroshot + priority 0.464 = majority floor; + banking77 contaminated; + raw margins not probabilities; + GH jev-haiku-benchmarking 404; + do not distill Jev as teacher of record (they distilled Haiku); + shas `56c9dba8` / `93e22fcf` / `4b6659d5`); + **measurement densifies PRIMARY** + ([Running-Dolphins/jev-bench](https://github.com/Running-Dolphins/jev-bench) + “0.9 is not one number”; ranking ≠ calibration; + banking77 0.8–0.9 stated 0.86 actual 0.73 over-confident *theirs*; + ≠ Praveenrajus/jev-bench ≠ fstandhartinger/jevbench; + Python MIT LICENSE SHA `891c70c6`; **0★**; HEAD `67d2ee42`; README SHA `ae2c5003`; size **1070**); + ([WallerChen/jev-measured](https://github.com/WallerChen/jev-measured) + $0.0000153–$0.0000226 vs circulating $0.0004 (~20×); + Score is 0..n-1 expectation not 0–1; + Noul has no confidence field; + TCP floor 198.8 ms; + type reliability is not a reason to choose Jev (json_schema 5/5); + gateway tax not one number; + Python MIT LICENSE SHA `0c563e25`; **0★**; HEAD `4a12dfb3`; README SHA `1c2c38ac`; size **37**); + ([RadRebelSam/jev-decision-lab](https://github.com/RadRebelSam/jev-decision-lab) + Function-only 5/8 vs hybrid 8/8; 4/8 without Jev; + 8 designed cases not conversion lift; + ≠ RadRebelSam/awesome-jev; + TS MIT LICENSE SHA `ae6f994c`; **0★**; HEAD `e6d6d42d`; README SHA `52a0071d`; size **128**); + ([jackojacko05/compare-jev-bigquery-ai-functions](https://github.com/jackojacko05/compare-jev-bigquery-ai-functions) + 200-row pilot Jev 86.5% 173/200 vs Gemini Flash-Lite 86.0% 172/200 vs Pro 87.0% 174/200 *theirs*; + not a ranking; + Jupyter MIT LICENSE SHA `0248c4b0`; **0★**; HEAD `cb9ef56b`; README SHA `d3db5d6c`; size **307**); + ([Trecto34/openjev-fighting-ring](https://github.com/Trecto34/openjev-fighting-ring) + NLI Tetris argmax P(entail)−P(contradict); + Python; license null; **0★**; HEAD `ac544f1e`; README SHA `b3fe2907`; size **1159**); + ([joshhu/jevtest](https://github.com/joshhu/jevtest) + 情緒測謊器; 1q 396ms / 30q 567ms; ±0.03; + 33q $0.000045 vs Gemini ~5× slower ~60× cost *theirs*; + ≠ realZachi/jevtest; + HTML; license null; **0★**; HEAD `6a4df41d`; README SHA `57d99325`; size **34**); + (hfspace:aahf/JevBenchmark + 8-example Jev vs GPT-5.6 Sol ~64× cost 5.4× latency *theirs*; + synthetic; no inference; ≠ JevBench v1.2 §78; + likes **1**; sha `36c28088`); + ([rhc98/awesome-jev](https://github.com/rhc98/awesome-jev) + Judged 3317 / listed 2560; Jev judges, code applies policy; + catalog ≠ endorsement; + TS; SPDX NOASSERTION LICENSE SHA `d491d714`; **1★**; HEAD `82898f25`; README SHA `e1f12080`; size **2529**); + ([AiPersonacademy/Awesome-jev-use](https://github.com/AiPersonacademy/Awesome-jev-use) + APA “microsecond policy / zero hallucination” overclaim; + CC0-1.0 LICENSE SHA `7ba0a23e`; **2★**; HEAD `29246f12`; README SHA `d506c11f`; size **124**); + **applied / theory placements** + ([erseco/questionator](https://github.com/erseco/questionator) + Client-side quiz; pointer from held docs; scanned-PDF warn; + CSP only api.typesafe.ai; + JS MIT LICENSE SHA `3ebdd11d`; **0★**; HEAD `b498f10b`; README SHA `4334c847`; size **1267**); + ([grgy078033/grill-jev](https://github.com/grgy078033/grill-jev) + Jev judges / agent reasons / user decides; + selecting an option is not permission to implement; + degraded fallback; + Python MIT LICENSE SHA `7c0e962d`; **1★**; HEAD `5ce6cbe8`; README SHA `da80efcc`; size **85**); + ([makefunstuff/jev-lsp](https://github.com/makefunstuff/jev-lsp) + pattern exact, judgement must clear floor; + no matching pattern → no model call; + not a correctness oracle; + $0.00022 vs chat $0.00306 *theirs*; + Rust MIT LICENSE SHA `fb5bc4a5`; **0★**; HEAD `bce3d8ed`; README SHA `c7ecbf8f`; size **1696**); + ([nozomi-koborinai/jev-spec](https://github.com/nozomi-koborinai/jev-spec) + Spec vs artifact remainder; + treating 0.85 as 85% / minProbability hard-gate as Harbor; + TS MIT LICENSE SHA `e43f732a`; **1★**; HEAD `86f14198`; README SHA `c0f29baa`; size **73**); + ([202620325-spec/Jev-LLM](https://github.com/202620325-spec/Jev-LLM) + VERIFY acquires discriminating evidence, never same-pool confidence-only rescoring; + fast/full/max are ceilings not sizes; + Solar writes, Jev chooses NEXT ACTION; + Python MIT LICENSE SHA `5d699c97`; **0★**; HEAD `da06b6d1`; README SHA `f371ba83`; size **150**). + SemIf 2186★ (+20 vs §109 2166); + jevlike 1038★ (+7 vs 1031); + TypeAR 14★ flat; + AnotiaWang 96★ (+1 vs 95); + yibie/awesome-jev 490★; + Laya likes 802 (was 783); + tracker likes 64 flat, lastModified UNCHANGED; + Laya present; Blackwood ABSENT; Archer still promised_not_landed. + 0★ HIGH still got a real card. Soft Noul ≠ hard safety. + `invented_signal: false`. + +- Hourly 0439 HIGH (`research/notes.md` §109): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + **HARD RULE:** do not reopen or amend PR #23 or #24 or #25 + or #26 (do not reopen or amend PR #23/#24/#25/#26. + merged #23 owns `notes.md` §105 / items 149–160 / + batch #88; merged #24 owns `notes.md` §106 / items 161–177 / + batch #89; merged #25 owns `notes.md` §107 / items 178–185 / + batch #90; merged #26 owns `notes.md` §108 / items 186–201 / + batch #91 — leave them alone). Never reopen merged + #7–**#26**. Do **not** re-fold 0345 / §108 / 0243 / §107 / + 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / §103 / + 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 / §99 / + 1740 / §98 / 1639 / §96 / gliner-native-runtime / §97 / + 1541 / §95 / jev-align *mechanism* / §93 / + jev-orderby-bench *six-gates* / §60 / JevBench v1.2 + *board* / §78 / openJev-verdict *claim-audit* / §71 / + yuki-oshio/mini-jev *93.25%* / §103. How-to-apply / mental + models / architecture / Harbor-jevals / toolbelt — not a + thin Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating + WANLI-256 64.5/60.2/52.0 as Harbor, label_mass as + correctness, Hub v1 family cards as shipping GGUF, + re-folding §108 27B/ternary as new, treating LoRA stub + cards as independent eval, hard-gating jevify n=307 ECE + 0.061 as “honest probabilities”, collapsing kushalpatil + into Mintzs/gulagala001/uspraveen, treating GH + kushalpatil07/jevify as an existing repo, collapsing + JulesHuisman into SivletLabs/4esv, quoting 32/32 synthetic + as production, hard-gating 0.8 evidence as proof, quoting + 40-band 0/10 without 7-band, treating /judge 0.5 as truth, + treating git-confess 11% as a person verdict, treating + paper-trader +12.40% as edge, treating Awesomejev + 656/38160 as eval, treating tracker likes as Archer + landing, using roadus2 spelling, pasting ultra_laya vs-Jev + as this-fork win, or treating jevgraph 100% gated as + production is soundness theater. jevify ECE 0.061 as a + hard gate, 0.8 evidence as proof, and paper-trader fills + as edge are the anti-patterns. Twenty-four HIGH clusters / + three themes: + **open reproduction densifies** + (hf:kushalpatil/jevify-gemma4-26b-a4b + Gemma-4 26B-A4B jevify classification+calibration; + Hub jevify merged LoRA ships weights; + PAWS 0.580/ece 0.288 is the weak cell; + kushalpatil/jevify-gemma4 ≠ Mintzs/jevify ≠ gulagala001/jevify ≠ uspraveen/Jevify; + GH kushalpatil07/jevify 404; + license gemma; likes **0**; sha `d4c0d1d4`); + (hf:kushalpatil/jevify-gemma4-26b-a4b-lora + LoRA adapter twin not independent eval; + license null; likes **0**; sha `ec4a3d22`); + (hf:kushalpatil/jevify-gemma4-e4b + Gemma-4 E4B jevify; + smaller E4B slightly better OOD ECE than 26B-A4B; + license gemma; likes **0**; sha `a6b5a716`); + (hf:kushalpatil/jevify-gemma4-e4b-lora + E4B LoRA stub card; license null; likes **0**; sha `cca1f55e`); + (hf:NicolaiMTLassen/bonzi-8b-v1-jev + bonzi Bonsai-8B v1 GGUF densify; WANLI-256 64.5% *theirs*; + rank #4 of 6; MIT; likes **0**; sha `588bc44e`); + (hf:NicolaiMTLassen/bonzi-1.7b-v1-jev + Bonsai-1.7B v1; WANLI-256 52.0% *theirs*; rank #6 of 6; + MIT; likes **0**; sha `48148bf9`); + (hf:NicolaiMTLassen/bonzi-4b-v1-jev + Bonsai-4B v1; WANLI-256 60.2% *theirs*; rank #5 of 6; + MIT; likes **0**; sha `d5545084`); + ([roadius2/ultra_laya](https://github.com/roadius2/ultra_laya) + roadus2 watch misspelling; lock roadius2/ultra_laya; + ultra_laya REVIEW defects; + default branch claude/laya-jev-review-gg5ppo; + Python Apache-2.0 LICENSE SHA `d9a10c0d`; **0★**; HEAD `0dff5bd2`; README SHA `da95bc05`; size **0** WITH CONTENTS); + **measurement densifies PRIMARY** + ([AHTOOOXA/jev-cyrillic-audit](https://github.com/AHTOOOXA/jev-cyrillic-audit) + XNLI EN 88.3% ECE 0.032 → RU 77.3% ECE 0.096; + Δ −11.0 pp [−14.2,−7.8]; ECE +0.063; + MASSIVE no detectable difference at n=600; + confidence is function of p_max (r=1.000); + Python MIT LICENSE SHA `f9b9193e`; **0★**; HEAD `7167894a`; README SHA `d36a36f5`; size **1302**); + ([JulesHuisman/jev-eval](https://github.com/JulesHuisman/jev-eval) + JulesHuisman/jev-eval scaffolding / README SHA c356a584 (was empty e69de29b); + JulesHuisman/jev-eval ≠ SivletLabs/jev-eval ≠ willkelly/jev-evaluation ≠ 4esv ≠ xxkuboxx ≠ onlyoneaman ≠ dayhaysoos/jevals; + Python; license null; **0★**; HEAD `96c2a110`; size **0** WITH CONTENTS); + ([tunahansahin897/what-is-jev](https://github.com/tunahansahin897/what-is-jev) + 947 repos scored; A 273 / B 302 / C 372; LLM rubric ≠ benches; + Python; SPDX NOASSERTION LICENSE SHA `790f3358`; **0★**; HEAD `71d53be2`; README SHA `3ae98c56`; size **6277**); + ([Mishkun/judge-jev](https://github.com/Mishkun/judge-jev) + judge-jev 0.5 still soft; + TS; license null; **0★**; HEAD `1bf495d3`; README SHA `3788ab6e`; size **65**); + **applied class placements** + ([LiuHao-1443/jev-table-tennis](https://github.com/LiuHao-1443/jev-table-tennis) + 7 bands 6/10 vs 40 bands 0/10; + Python MIT LICENSE SHA `777dc3d6`; **1★**; HEAD `0224c21c`; README SHA `3a54bc93`; size **0** WITH CONTENTS); + ([laguagu/jev-evidence-lab](https://github.com/laguagu/jev-evidence-lab) + source receipts + confidence slider re-policy without re-inference; + 32/32 synthetic is smoke not production; + Python MIT LICENSE SHA `e5ede121`; **0★**; HEAD `a9669e94`; README SHA `1be53f55`; size **0** WITH CONTENTS); + ([hemanth/hfjev](https://github.com/hemanth/hfjev) + classify HF datasets across typed semantic dimensions; + Python MIT LICENSE SHA `0ac06d77`; **1★**; HEAD `6f2aa501`; README SHA `20651a25`; size **48**); + ([akash-kamat/jev-llm](https://github.com/akash-kamat/jev-llm) + pointer-not-generator 400 human-authored responses; + JS; license null; **0★**; HEAD `194db1a6`; README SHA `c42907a6`; size **54**); + ([chenmingtang830/jevgraph](https://github.com/chenmingtang830/jevgraph) + proposed ≠ authorized; FewRel 160: Jev 85.0% vs lexical 13.125%; + gated 100% (95/95) coverage 59.375%; + Python Apache-2.0 LICENSE SHA `57bc88a1`; **0★**; HEAD `7a6f7d05`; README SHA `bda2ed37`; size **125**); + ([Towow-ai/jpp](https://github.com/Towow-ai/jpp) + J++ composable semantic computation language; + Python MIT LICENSE SHA `b1a46fbc`; **3★**; HEAD `14d77789`; README SHA `0a82d97f`; size **1071**); + ([whyashthakker/awesome-jev-use-cases](https://github.com/whyashthakker/awesome-jev-use-cases) + No benchmark winner is claimed; + HTML MIT LICENSE SHA `13b5a2a6`; **3★**; HEAD `74583663`; README SHA `a6d96b87`; size **426**); + ([dog-last/awesome-jev](https://github.com/dog-last/awesome-jev) + phishing: naive 62.6% vs regex 91.8%; 5-atomic + LR 95.0% *theirs*; + Python MIT LICENSE SHA `cd31b9aa`; **1★**; HEAD `206fdcab`; README SHA `1ca65f69`; size **85**); + ([shinshin86/jev-aituber-tension-sample](https://github.com/shinshin86/jev-aituber-tension-sample) + AITuber tension ±15; + TS MIT LICENSE SHA `8b20d89e`; **0★**; HEAD `8c0a8ffb`; README SHA `06536d19`; size **105**); + ([shinpr/jev-reranker](https://github.com/shinpr/jev-reranker) + README npm global; repo is Rust; + Rust MIT LICENSE SHA `4306a712`; **1★**; HEAD `731deba3`; README SHA `e2818c0e`; size **163**); + ([AHTOOOXA/git-confess](https://github.com/AHTOOOXA/git-confess) + git-confess code owns counting/blame/ratio; + httpx exhibit 11% (13/119) *theirs*; + Python MIT LICENSE SHA `f9b9193e`; **0★**; HEAD `54cd2849`; README SHA `764076a6`; size **3**); + ([waterme7on/jev-paper-trader](https://github.com/waterme7on/jev-paper-trader) + 90d trend +12.40% vs random +12.75% vs BH +41.71%; + 5m win rate 25%; + JS; license null; **0★**; HEAD `73662f79`; README SHA `4eded35a`; size **145**). + Awesomejev 656 entries / 38,160 stars; + WANLI-256 64.5% / 60.2% / 52.0% *theirs*; + rank #4 / #5 / #6 of 6; + tracker likes 64 (+4) lastModified UNCHANGED; + Laya present; Blackwood ABSENT; Archer still promised_not_landed. + 0★ HIGH still got a real card. Soft Noul ≠ hard safety. + `invented_signal: false`. + +- Hourly 0345 HIGH (`research/notes.md` §108): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + **HARD RULE:** do not reopen or amend PR #23 or #24 or #25 (merged + #23 owns `notes.md` §105 / items 149–160 / batch #88; + merged #24 owns `notes.md` §106 / items 161–177 / batch #89; + merged #25 owns `notes.md` §107 / items 178–185 / batch #90 — + leave them alone). Never reopen merged + #7–**#25**. Do **not** re-fold 0243 / §107 / 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / + §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 + / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime + / §97 / 1541 / §95 / jev-align *mechanism* / §93 / + jev-orderby-bench *six-gates* / §60 / JevBench v1.2 + *board* / §78 / openJev-verdict *claim-audit* / §71 / + yuki-oshio/mini-jev *93.25%* / §103. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating + WANLI-256 as Harbor, label_mass as correctness, Hub + family cards as shipping weights, collapsing stock Q1_0 + into §107 Q2_0 gibberish, treating the HF twin as a + weights drop, quoting 93.25% as family-disjoint, + hard-gating injection ECE 0.058 as “Jev is calibrated”, + quoting AbstentionBench rank 1 without question-asymmetry / + 2025 field, treating dspy-bench as a quality claim, + hard-gating pdf-race 12/12 as pipeline equality, treating + atlas listed counts as eval, hard-gating flopcheck + composite as truth, treating calibration-lab 40-row T as + production, treating CLEVR joint 0% as “vision Jev works”, + treating ONNX 63/63 as ECE, treating openkev T as + transferable, or treating select_threshold inf as a bug + is soundness theater. Injection ECE 0.058 as a hard gate + and 40-row T as production are the anti-patterns. Sixteen + HIGH clusters / two themes: + **open reproduction densifies** + (hf:NicolaiMTLassen/bonzi-27b-v2-jev + ternary Bonsai 27B v2 open-bonzi GGUF family densify; + Hub still does not ship weights; WANLI-256 74.6% *theirs*; + ternary still needs PrismML fork; MIT; likes **0**; sha `c4d21b74`); + (hf:NicolaiMTLassen/bonzi-8b-ternary-v1-jev + ternary Bonsai 8B open-bonzi GGUF; WANLI-256 65.2% *theirs*; + MIT; likes **0**; sha `47b66187`); + (hf:NicolaiMTLassen/bonzi-27b-v1-jev + Bonsai 27B v1 open-bonzi GGUF; Bonsai 1 27B Q1_0 runs on stock llama.cpp; + WANLI-256 71.1% *theirs*; MIT; likes **0**; sha `2fb8061a`); + (hf:mizchi/laya-multilingual-onnx + Laya multilingual ONNX WebGPU typed-decisions port; + 63/63 selected answers / 5.1e-4 CPU / 1.2e-2 WebGPU; + ships model.onnx 646.9MB; apache-2.0; likes **0**; sha `b6314ec9`); + (hf:IamBusy/OpenJev-Vision + OpenJev Vision image classification + uncertainty; + CLEVR-4 held-out joint 0%; license other; likes **0**; sha `8cf6cbd3`); + (hf:heman10x/openJev-verdict-2.0 + twin tokenizer-only; no 149.6M weights; likes **5** ≠ GH **107★**; + apache-2.0; sha `794d5e0d`); + (hfdataset:IamBusy/OpenJev-Vision-Research-v0.1 + 12,832 images; 294,912 derived targets not independent samples; + license other; likes **0**; sha `43e49184`); + ([UpHash-Network/mini-jev](https://github.com/UpHash-Network/mini-jev) + UpHash-Network/mini-jev is yuki-oshio transfer; + residual-head 9,222-param decreased 73/96→67/96; + Python MIT LICENSE SHA `79e2cff7`; **0★**; HEAD `52fbae12`; README SHA `363441b6`; size **38388**); + **measurement densifies PRIMARY** + ([ASEVlad/jev-injection-bench](https://github.com/ASEVlad/jev-injection-bench) + jev-injection-bench 11,900 labelled prompts; + Jev best ranking / Haiku better ECE 0.021 vs 0.058; + 0.5–0.9 band is where Jev's numbers do not mean what they say; + Prompt wording moves panic 28%; + Python MIT LICENSE SHA `0dab34c6`; **0★**; HEAD `c0d0f25d`; README SHA `1b28498c`; size **107**); + ([manojlds/jev-dspy-bench](https://github.com/manojlds/jev-dspy-bench) + Jev agreement is similarity, never ground truth; + no aggregate quality grade or merge gate; + manojlds/jev-dspy-bench ≠ dspachos/jev-dspy ≠ jmanhype/jev-dspy-lab; + Python Apache-2.0 LICENSE SHA `261eeb9e`; **0★**; HEAD `d8c68d72`; README SHA `16100deb`; size **0** WITH CONTENTS); + ([sshariqali/jev-abstentionbench](https://github.com/sshariqali/jev-abstentionbench) + AbstentionBench-on-Jev rank 1 of 20 vs 2025 field; + question-asymmetry; forward-looking 0.465 never extreme; + Python; README MIT / GitHub SPDX NOASSERTION LICENSE SHA `7f6fb3ba`; **0★**; HEAD `f5c0c068`; README SHA `26393f0f`; size **0** WITH CONTENTS); + ([misakaikato/openkev](https://github.com/misakaikato/openkev) + openkev calibration layer not a runtime; + ECE vs coverage independent; select_threshold returns inf; + escalation catches uncertainty not ignorance; + misakaikato/openkev ≠ jaredpalmer/kev; + Python MIT LICENSE SHA `8c3231cb`; **0★**; HEAD `babcab1c`; README SHA `974d97f9`; size **484**); + ([goodrahstar/pdf-race](https://github.com/goodrahstar/pdf-race) + pdf-race Docling→Jev vs Gemini; parser owns the wall clock; + 12/12 tie is a tie; titles selected not generated; + JS MIT LICENSE SHA `182ac1d5`; **0★**; HEAD `1c687fc6`; README SHA `05730787`; size **0** WITH CONTENTS); + ([ZeroX-01/jev-atlas](https://github.com/ZeroX-01/jev-atlas) + ZeroX-01/jev-atlas ≠ Zaious/jev-capability-atlas ≠ gorock007/jev-atlas; + catalog not endorsement; + JS; license null; **0★**; HEAD `bc94bf50`; README SHA `8dd3d308`; size **0** WITH CONTENTS); + ([samyakjain0606/jev-is-here](https://github.com/samyakjain0606/jev-is-here) + flopcheck 16 calibrated tweet judgments; mechanical tells in code; + TS; license null; **0★**; HEAD `eb0de0ba`; README SHA `cefb8f96`; size **0** WITH CONTENTS); + (hfspace:BunsDev/laya-calibration-lab + Laya calibration lab Gradio MCP; T never changes argmax; + confidence ≠ top-label p; easy probe set refused; + 40–48 rows too small to ship T; apache-2.0; likes **0**; sha `a3fc13ba`). + 0★ HIGH still got a real card. Soft Noul ≠ hard safety. + `invented_signal: false`. + +- Hourly 0243 HIGH (`research/notes.md` §107): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + **HARD RULE:** do not reopen or amend PR #23 or #24 (merged + #23 owns `notes.md` §105 / items 149–160 / batch #88; + merged #24 owns `notes.md` §106 / items 161–177 / batch #89 — + leave them alone). Never reopen merged + #7–**#24**. Do **not** re-fold 0145 / §106 / 0042 / §105 / 2340 / §104 / 2246 / + §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 + / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime + / §97 / 1541 / §95 / jev-align *mechanism* / §93 / + jev-orderby-bench *six-gates* / §60 / JevBench v1.2 + *board* / §78 / openJev-verdict *claim-audit* / §71. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating a + 0.85 figure FAST_PATH as a proof, 100/100 easy T/F as + Harbor, label_mass as correctness, 77.10% as beating + Jev, collapsing ticket-router into Sarath, collapsing + jev-bench into jevbench, collapsing DeBERTa ONNX into + the Qwen scorer ONNX, collapsing Hub/GH Nicolai + spellings, treating 151M vs 149.6M as two models, + quoting 62.3% cost save without the 4.5pp miss, + treating mock keyword as production, hard-gating + ticket 0.6 as safety, treating a study mock as a live + API, treating stock llama.cpp Q2_0 as working, or + re-folding §71 as a beat is soundness theater. + Figure-router 0.85 hard-gate and 100/100 easy T/F as + Harbor are the anti-patterns. Eight HIGH clusters / + three themes: + **measurement densifies PRIMARY** + ([erendikmenn/jev-llm-router-benchmark](https://github.com/erendikmenn/jev-llm-router-benchmark) + Benchmark-driven Jev router and judge; cheap alone is not success; + Jev does not write, sum prices, or claim accuracy %; + Sol 94.2 / Luna 83.9 / Jev path 89.7; + 19.2% Sol / 62.3% cost save / 4.5pp miss of 2pp non-inferiority; + p50 latency worse than Sol due to routing overhead; + erendikmenn/jev-llm-router-benchmark ≠ jev-rag-benchmark ≠ ryantsai/jev-llm-router; + Python MIT LICENSE SHA `99916677`; **0★**; HEAD `f44ef450`; README SHA `df3687e5`; size **0** with contents); + ([aesaganda/jev-ticket-router](https://github.com/aesaganda/jev-ticket-router) + Express + node:sqlite; mock and Jev decision engines; + previous_ticket_count >= 3 is code; MIN_CONFIDENCE 0.6 still soft; + substring false positives; + aesaganda/jev-ticket-router ≠ SarathChandraBellam/jev-vs-llm-ticket-router; + JS; license null; **0★**; HEAD `ecf00046`; README SHA `41c0b50f`; size **0** with contents); + ([hoangngochuong24947-gif/jev-figure-router](https://github.com/hoangngochuong24947-gif/jev-figure-router) + Universal Figure & Diagram Router; confidence ≥ 0.85 hard-gate is theater; + generative AI banned from scientific plots; six visual branches; + Python MIT LICENSE SHA `62120c23`; **1★**; HEAD `4c51b1af`; README SHA `11d5302b`; size **340**); + (hfdataset:Praveenrajus/jev-bench + human-labeled (state, question, label); 166,054 rows / 22 configs; + soft_label for human uncertainty; + Praveenrajus/jev-bench ≠ fstandhartinger/jevbench; + license other; likes **0**; sha `c9c3032c`); + **open reproduction class ports** + (hf:NicolaiMTLassen/open-bonzi-jev + ternary bonsai System One GGUF; openjev's mechanism, Bonsai's weights; + Hub does not ship weights; 100/100 easy T/F is not Harbor; + label_mass ≠ correctness; stock llama.cpp Q2_0 silently gibberish; + NicolaiMTLassen/open-bonzi-jev ≠ NicolaiLassen; + MIT; likes **0**; sha `09240156`; + GH companion NicolaiLassen/open-bonzi-jev Python MIT; **0★**; HEAD `1c2508fd`; LICENSE SHA `46d740e4`; size **0** with contents); + (hf:onnx-community/open-jev-deberta-v3-large-ONNX + transformers.js DeBERTa ONNX; source:com-kotobalabs/open-jev-deberta-v3-large; + temperature 1.05; AutoModel from_pretrained works; + onnx-community/open-jev-deberta-v3-large-ONNX ≠ system-one-qwen3.5-4b-scorer-ONNX; + apache-2.0; likes **0**; sha `3bc2553b`); + ([Heman10x-NGU/openJev-verdict-2.0](https://github.com/Heman10x-NGU/openJev-verdict-2.0) + 107★ densify; GH 151M vs README 149.6M; PR #1 now closed unmerged; + do not re-fold §71 claim-audit as a beat; + Python; README Apache-2.0 / GitHub SPDX NOASSERTION; **107★**; HEAD `a458733c`; README SHA `05ca75af`; size **14728**); + **study densification** + ([wjdjdakf17/jev-study](https://github.com/wjdjdakf17/jev-study) + typed decisions, RLCD, confidence-gated routing; structured ≠ correct; + mock not live API; 26 tests; wjdjdakf17/jev-study ≠ baekenough/jev-study; + TypeScript MIT LICENSE SHA `00bbf5f9`; **0★**; HEAD `24b5d7d7`; README SHA `4ca30c93`; size **34**; default **master**). + 0★ HIGH still got a real card. Soft Noul ≠ hard safety. + `invented_signal: false`. + +- Hourly 0145 HIGH (`research/notes.md` §106): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + **HARD RULE:** do not reopen or amend PR #23 (merged + #23 owns `notes.md` §105 / items 149–160 / batch #88 — + leave it alone). Never reopen merged + #7–**#23**. Do **not** re-fold 0042 / §105 / 2340 / §104 / 2246 / + §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 + / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime + / §97 / 1541 / §95 / jev-align *mechanism* / §93 / + jev-orderby-bench *six-gates* / §60 / JevBench v1.2 + *board* / §78. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating a + Jevify two-liner as a checkpoint, collapsing Jevify + into Mintzs or gulagala001, pasting Colvin as hyusi, + treating JSON parse as a Noul, quoting 46x / 91.1% / + 92.58% / 84.8 as class ceilings, treating + classifier.dev #1 as a better model, re-folding §78 + as new, letting Jev do budget arithmetic, treating + XState as Jev, pasting catalog ★ as eval, treating + SPLADE as TypeSafe Jev, or treating an empty + consistency Space as a win is soundness theater. + Jevify stub and JSON-parse-as-Noul are the + anti-patterns. Seventeen HIGH clusters / four themes: + **architecture probes PRIMARY** + ([uspraveen/Jevify](https://github.com/uspraveen/Jevify) + Turn any open LLM into System-One Jev; + uspraveen/Jevify ≠ Mintzs/jevify ≠ gulagala001/jevify; + Jevify-any-LLM architecture probe; description-only stub / size 0; + license null; **0★**; HEAD `0f29d783`; README SHA `32608473`); + ([Ruivalim/exu-base](https://github.com/Ruivalim/exu-base) + Train encoder-only calibrated decision models from a task sentence; + Exu is a toolkit, not a method; strictly proper scoring rule; Pre-alpha; + Python MIT LICENSE SHA `336cde4c`; **1★**; HEAD `7288cdca`; README SHA `45244838`; size **180**); + ([Colvin0315/MiniSystemOne](https://github.com/Colvin0315/MiniSystemOne) + scratch-trained calibrated decision model; typed Q → probability dists; + Colvin0315/MiniSystemOne ≠ hyusi2003/MiniSystemOne; + no published weights download URL; 90.5 seconds / 29.2% pipeline evidence; + p_i/p_j independent of other candidates; + Apache-2.0 LICENSE SHA `261eeb9e`; **0★**; HEAD `d7f9f803`; README SHA `a5b0d2fd`; size **814**); + ([scienthoon/luce](https://github.com/scienthoon/luce) + Recipe for calibrated decision models — small model out; + init → synth → train → eval → serve; 91.1 % / ECE 0.022 *theirs*; + Jev zero-shot 75.1; Apache-2.0 LICENSE SHA `d6456956`; **1★**; HEAD `8072b97d`; README SHA `5fbe226d`; size **10833**); + ([RichardoMrMu/jev-mini](https://github.com/RichardoMrMu/jev-mini) + Put Jev's three headline claims on trial; 0.5B local GPU; + 46x speedup / accuracy identical; ECE 0.624 sentiment catastrophe; + bigger model worse calibration; + RichardoMrMu/jev-mini ≠ yuki-oshio/mini-jev ≠ r-ms/mini-jev; + Python MIT LICENSE SHA `cbdf24bb`; **0★**; HEAD `5adef5dc`; README SHA `154ae8b0`; size **0** with contents); + ([tapsin/jev-local](https://github.com/tapsin/jev-local) + System-1 decision engine for local LLMs; structured choices only; + JSON parse of generated text ≠ Noul; TypefAI JEV / Journal Entry Voucher; + tapsin/jev-local ≠ us/jev-local ≠ Argos1111/jev_local; + Python; license null; **0★**; HEAD `96aac2a1`; README SHA `4d728cbf`; default **master**; size **0** with contents); + **measurement densifies** + ([goya4140/jev-reward-model-evaluation](https://github.com/goya4140/jev-reward-model-evaluation) + Jev 1.13 reward-model eval across 8 benchmark tracks; + 40,940 examples / 0 API errors; RewardBench v1 92.58%; Precise IF 50.63%; + Python MIT LICENSE SHA `4588ffe5`; **0★**; HEAD `f16a06d1`; README SHA `603587fd`; size **0** with contents); + ([SarathChandraBellam/jev-vs-llm-ticket-router](https://github.com/SarathChandraBellam/jev-vs-llm-ticket-router) + Jev vs LLM support-ticket routing; Scaffolding in progress; + license null; **0★**; HEAD `0318ad72`; README SHA `8223dd6f`; default **master**; size **0**); + ([Shilin237/jev-vs-llm-cost](https://github.com/Shilin237/jev-vs-llm-cost) + static + live decision bench; TypeSafe's own published benchmark; + illustrative simulations, not live API calls; + HTML; license null; **0★**; HEAD `2d4bd6a9`; README SHA `f4e98bbf`; default **master**; size **0**); + ([fstandhartinger/jevbench](https://github.com/fstandhartinger/jevbench) + JevBench v1 — smart/cheap/fast/reliable; I/C/S/K 25% geometric mean; + classifier.dev fast tier 84.8 is Jev behind its own API; + do not re-fold §78 v1.2 board as new; Laya (421M) 70.1 now on board; + Python MIT LICENSE SHA `ebdd5738`; **6★**; HEAD `c7ab99f5`; README SHA `8fe07c41`; size **8907**); + ([AIGNLAI/ReflexRoute](https://github.com/AIGNLAI/ReflexRoute) + Zero-shot/few-shot LLM routing; hard budget filter before Jev; + Jev never asked to perform budget arithmetic; + Python MIT LICENSE SHA `1be8631e`; **1★**; HEAD `5f475d80`; README SHA `6a1ae694`; size **0** with contents); + ([priyankark/jev-state](https://github.com/priyankark/jev-state) + Jev judges the next state, XState enforces transitions; + simulation uses synthetic keyword fixtures; + TypeScript MIT LICENSE SHA `78745387`; **0★**; HEAD `c73aef65`; README SHA `15410dbe`; size **0** with contents); + (hfspace:mjyoke1111/jev-consistency-benchmark + Consistency benchmark Space; This Space contains no benchmark result yet; + 12-case plumbing fixture; apache-2.0; likes **0**; lastModified `2026-09-20T05:29:45Z`); + **catalog gravity** + ([v-modal/awesome-jev-tools](https://github.com/v-modal/awesome-jev-tools) + catalog gravity; ★339 live REST; curation is not endorsement; + license null; **339★**; HEAD `f117e0c3`; README SHA `8ea9a669`; size **83**); + ([RadRebelSam/awesome-jev](https://github.com/RadRebelSam/awesome-jev) + crawler-maintained directory; Daily GitHub + npm sweep, human-merged; + RadRebelSam/awesome-jev ≠ AnotiaWang ≠ yibie ≠ cobanov ≠ logicrw ≠ v-modal; + SPDX NOASSERTION LICENSE SHA `2aa7fe23`; **0★**; HEAD `27629954`; README SHA `21ec8d51`; size **919**); + **HF class ports** + (hf:rdxtremity/jev-reranking + HF peft SPLADE/BGE reranker; rdxtremity/jev-reranking ≠ carlaiau/jev-reranking; + query-side encoders, not a Jev replica; apache-2.0; likes **0**; sha `cae796ea`); + (hf:onnx-community/system-one-qwen3.5-4b-scorer-ONNX + ONNX System One Qwen3.5-4B scorer; source:pngwn/system-one-qwen3.5-4b-scorer; + CC-BY-NC-4.0; temperature 1.75; transformers.js AutoModel cannot load this graph; + likes **0**; sha `fa0bed22`). + 0★ HIGH still got a real card. Soft Noul ≠ hard safety. + `invented_signal: false`. + +- Hourly 0042 HIGH (`research/notes.md` §105): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#22**. Do **not** merge from + merged **#22** (hourly 2340 / §104). Do **not** re-fold 2340 + / §104 / 2246 / §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / + 1843 / §99 / 1740 / §98 / 1639 / §96 / + gliner-native-runtime / §97 / 1541 / §95 / jev-align + *mechanism* / §93. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating + displayed p as proof, schema-valid ordinary-model JSON + as a calibrated Noul, collapsing gulagala001/jevify into + Mintzs/jevify, quoting 40.3% without the constant-answer + baseline, hard-gating fastloop confidence as safety, + treating fan-out 0.0000 sd as universal determinism, + treating 2.95× RL as Harbor, hard-gating Laya cascade + 0.60, collapsing locate into decide, treating snake + points as intelligence, treating jevcheck as a + correctness proof, treating `no_direct_evidence` as + merge-safe, or treating a theme Choice as a music-theory + certificate is soundness theater. Hard-gating a soft + Noul as safety is the anti-pattern. Unique consecutive fragments: structured probability readouts; distribution > argmax; Noul 0.5 midpoint; score is expectation not integer; bare HTTP not SDK; Arohtea/jev-readout; Jev-style Choice/Score/Noul from ordinary models; optional DSH plugin; schema-valid ≠ calibrated; gulagala001/jevify ≠ Mintzs/jevify; Laya RLCD benchmark; 40.3% below constant-answer; open-weight measurement; mourad-ghafiri/laya-rlcd-benchmark ≠ yibie/laya-jev-lab; cheap fail-open semantic edge; second signal not sole; FastLoopError catch; SupremeDreamZ/jev-fastloop ≠ jev-ultrafast; asking more questions in one call; 0.980 at every N; nearly not fully deterministic; TheWebDevel/jev-fanout; Qwen3-VL perception + Jev decisions train RL; 0 model calls at deployment; VLM alone 1.7 vs +Jev 4.4; harneet2512/reflexrl ≠ khordoo/jev-reflex-autonomy-lab; independent Jev API vs Laya; cascade 0.60 matches 78% at 1.8×; noul facts not judgements; yibie/laya-jev-lab ≠ dairui1/jev-lab ≠ BrendanH18/jev-lab; GLiNER vs GLiFormer vs Laya vs Jev; extractors ≠ decision engines; Laya dict-instructions collapse 58.3%; umstek/zero-shot-ie-bench; decisions-per-minute & cost; 204 moves vs 73; throughput not intelligence; angelgalvisc/snake-arena-jev-vs-llms ≠ vtrivedy/jev-plays-games; behavioral contracts; pin expectations eval upgrades; raw 0.94 is not a release; sathariels/jevcheck ≠ dayhaysoos/jevals ≠ SivletLabs/jev-eval; evidence-linked dependency upgrade; Jev never generates filenames; no_direct_evidence ≠ safe to merge; GaneshVG18/upgrade-radar ≠ LYchoon/paper-radar-jev; discography theme/mood/complexity; five atomic questions one call; lirantal/discoprint. Twelve HIGH + clusters: **structured probability readouts** + ([Arohtea/jev-readout](https://github.com/Arohtea/jev-readout) + PRIMARY; structured probability readouts; distribution > + argmax; Noul 0.5 midpoint; score is expectation not + integer; bare HTTP not SDK; JavaScript; license null; + **0★**; HEAD `6f1e5900`; README SHA `67ee1e96`); + **ordinary-model Jev-shape** + ([gulagala001/jevify](https://github.com/gulagala001/jevify) + Jev-style Choice/Score/Noul from ordinary models; + optional DSH plugin; schema-valid ≠ calibrated; + gulagala001/jevify ≠ Mintzs/jevify; JavaScript MIT; + **0★**; HEAD `3d3e904a`; README SHA `0e2f8536`); + **open-weight Laya measurement** + ([mourad-ghafiri/laya-rlcd-benchmark](https://github.com/mourad-ghafiri/laya-rlcd-benchmark) + Laya RLCD benchmark; 40.3% below constant-answer; + open-weight measurement; + mourad-ghafiri/laya-rlcd-benchmark ≠ yibie/laya-jev-lab; + Python; license null; **0★**; HEAD `3401ff26`; README SHA + `d8d4859e`); + **cheap fail-open semantic edge** + ([SupremeDreamZ/jev-fastloop](https://github.com/SupremeDreamZ/jev-fastloop) + cheap fail-open semantic edge; second signal not sole; + FastLoopError catch; SupremeDreamZ/jev-fastloop ≠ + jev-ultrafast; Python MIT; **0★**; HEAD `1157841a`; + README SHA `053a0885`; GitHub size **12** (relock; was + **0** with contents)); + **fan-out measurement** + ([TheWebDevel/jev-fanout](https://github.com/TheWebDevel/jev-fanout) + asking more questions in one call; 0.980 at every N; + nearly not fully deterministic; Python MIT; **0★**; HEAD + `b30aaadc`; README SHA `394b2e1e`; GitHub size **206** + (relock; was **0** with contents)); + **VLM+Jev RL teacher** + ([harneet2512/reflexrl](https://github.com/harneet2512/reflexrl) + Qwen3-VL perception + Jev decisions train RL; 0 model + calls at deployment; VLM alone 1.7 vs +Jev 4.4; + harneet2512/reflexrl ≠ khordoo/jev-reflex-autonomy-lab; + Python MIT; **0★**; HEAD `aa36be84`; README SHA + `e6ff13cd`; default master; GitHub size **749** (relock; + was **0** with contents)); + **independent Jev API vs Laya** + ([yibie/laya-jev-lab](https://github.com/yibie/laya-jev-lab) + independent Jev API vs Laya; cascade 0.60 matches 78% at + 1.8×; noul facts not judgements; yibie/laya-jev-lab ≠ + dairui1/jev-lab ≠ BrendanH18/jev-lab; Python MIT; **0★**; + HEAD `30ba64dc`; README SHA `57bd1832`); + **locate vs decide** + ([umstek/zero-shot-ie-bench](https://github.com/umstek/zero-shot-ie-bench) + GLiNER vs GLiFormer vs Laya vs Jev; extractors ≠ + decision engines; Laya dict-instructions collapse 58.3%; + Python MIT; **0★**; HEAD `8770b16b`; README SHA + `d69dc96a`); + **throughput arena** + ([angelgalvisc/snake-arena-jev-vs-llms](https://github.com/angelgalvisc/snake-arena-jev-vs-llms) + decisions-per-minute & cost; 204 moves vs 73; throughput + not intelligence; angelgalvisc/snake-arena-jev-vs-llms ≠ + vtrivedy/jev-plays-games; Python MIT; **0★**; HEAD + `985a1c70`; README SHA `0db6f528`); + **behavioral contracts** + ([sathariels/jevcheck](https://github.com/sathariels/jevcheck) + behavioral contracts; pin expectations eval upgrades; + raw 0.94 is not a release; sathariels/jevcheck ≠ + dayhaysoos/jevals ≠ SivletLabs/jev-eval; Python MIT; + **0★**; HEAD `fc49c795`; README SHA `5c832f81`); + **evidence-linked upgrade review** + ([GaneshVG18/upgrade-radar](https://github.com/GaneshVG18/upgrade-radar) + evidence-linked dependency upgrade; Jev never generates + filenames; no_direct_evidence ≠ safe to merge; + GaneshVG18/upgrade-radar ≠ LYchoon/paper-radar-jev; + TypeScript MIT; **0★**; HEAD `e438f9bd`; README SHA + `c32f7d18`); + **knowledge-work discography** + ([lirantal/discoprint](https://github.com/lirantal/discoprint) + discography theme/mood/complexity; five atomic questions + one call; TypeScript Apache-2.0; **0★**; HEAD + `a9d3294f`; README SHA `9a64f473`). Soft Noul ≠ hard + safety. 0★ HIGH still got a real card (all twelve + **0★**). Census **live REST pulse** (independent review + relock after `0558f7d`): SemIf **2069★** (+22 vs §104 **2047**); jevlike + **1022★** (+4 vs §104 **1018**); TypeAR **12★** **flat**. + Tracker likes **59** (+3 vs §104 **56**); lastModified + `2026-09-20T04:29:16Z` **UNCHANGED** vs §104. Hub + `archerhume/4rcherhume` HTTP **401**. Archer still NOT + landed. Laya Hub likes **705**. Blackwood Hub likes **2** + gated manual — user census absent from tracker. Awesomejev + 561/27007 user-provided (≠ AnotiaWang/awesome-jev + **92★** ≠ yibie/awesome-jev **450★**). Qwen3.8-27B ≠ Archer (likes **15796**). + `invented_signal: false`. Do not copy keys / `npm` / + `pip` / `npx` / `uv` / `.env`. No wrapper. No invented + metrics. Relock: Arohtea HEAD `03734993`→`6f1e5900` / + README `76a99fe5`→`67ee1e96`; upgrade-radar HEAD + `1e91eb03`→`e438f9bd` / README `b7603c01`→`c32f7d18`; + discoprint HEAD `c2a3d47e`→`a9d3294f` / README + `392efd29`→`9a64f473`; yibie/laya-jev-lab size **49** + (was **0** with contents); umstek size **50** (was + **33**); jev-fastloop size **12** (was **0** with + contents); jev-fanout size **206** (was **0** with + contents); reflexrl size **749** (was **0** with + contents); upgrade-radar size **908** (was **0** with + contents); gulagala001/jevify size **145** (was **0** + with contents); mourad size **164** (was **0** with + contents); discoprint size **239** (was **0** with + contents); Arohtea size **44** (was **0** with + contents). Design claims unchanged. +- Hourly 2041 HIGH (`research/notes.md` §101): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#18**. Do **not** re-fold 1943 + / §100 / 1843 / §99 / 1740 / §98 / 1639 / §96 / + gliner-native-runtime / §97 / 1541 / §95 / jev-align + *mechanism* / §93. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating a + zero binary name gap as a fairness certificate, letting + Jev emit the verdict string, pasting PrismNLI's lead + without the contamination caveat, treating J7 pass as + safe to obey, treating a receipt as authorization, or + hard-gating confidence ≥0.95 is soundness theater. Nine + HIGH clusters: **resume-screening bias audit** + ([natemoo-re/bias-bench](https://github.com/natemoo-re/bias-bench) + PRIMARY; name×resume factorial independent Nouls; + callback determined by resume quality; mean-probability + name gaps operationally negligible; natemoo-re/bias-bench + ≠ BBQ; JavaScript; license null; **0★**; HEAD + `fe2f2535`; README SHA `a1c3e604`); **MCDA panel + code-owned verdict** + ([austindixson/planalyzer](https://github.com/austindixson/planalyzer) + Plan/PRD panel → code-owned pass|review|block; + cheerleading out of scope; Python MIT; **0★**; HEAD + `39fc161f`; README SHA `6e4d8da3`); **EU cost-aware + routing** + ([cannacre8ive/switchboard-ai](https://github.com/cannacre8ive/switchboard-ai) + cost-aware multi-model routing/escalation; decide vs do; + successful-task cost; JavaScript MIT; **1★**; HEAD + `5cae9d1c`; README SHA `872de837`; package **0.4.0**); + **frozen-protocol class bake-off** + ([elcronos/jev-vs-open-decision-models](https://github.com/elcronos/jev-vs-open-decision-models) + TypeSafe Jev vs PrismNLI vs Laya; contamination caveat; + Python; license null; **0★**; HEAD `b61e6cfc`; README + SHA `b7256888`); **VOI admission** + ([cvsgireesh/jevusher](https://github.com/cvsgireesh/jevusher) + context-window admission control; fail polarity per + lens; on small inputs lenses lose money; TypeScript MIT; + **0★**; HEAD `d830d344`; README SHA `428a4a59`); + **Leveson control plane** + ([MokiMeow/jev-fabric](https://github.com/MokiMeow/jev-fabric) + typed decision control plane; receipt ≠ authorization; + historical-v0 zero retained cases; Apache-2.0; **0★**; + HEAD `95b9a4f3`; README SHA `f485dbdd`); **scoring + economics** + ([jose-troche/live-rubric](https://github.com/jose-troche/live-rubric) + live 15-dim typed rubric re-score per pause; OpenJev/Codiv + ≠ TypeSafe hosted; ~$0.000004 desc / ~$0.000006 README; + TypeScript; license null; **0★**; HEAD `db8da8db`; + README SHA `4a0be084`); **pre-registered calibration + science** + ([willkelly/jev-evaluation](https://github.com/willkelly/jev-evaluation) + 28 predictions before data; 123,805 requests; confidence + does not track ignorance; polite injection 65% / crude + 0%; Python MIT; **0★**; HEAD `c168c093`; README SHA + `2d66ac22`; rh-guard owns injection); **class + infrastructure SDK** + ([nshkrdotcom/system_one_sdk](https://github.com/nshkrdotcom/system_one_sdk) + provider-neutral Elixir/BEAM Noul/Choice/Score SDK; + GitHub desc provider-neutral / README TypeSafe-first; + Elixir MIT; **0★**; HEAD `c2a522ee`; README SHA + `c117b4c4`; mix **0.5.0**). Soft Noul ≠ hard safety. + 0★ HIGH still got a real card (switchboard now **1★** + live REST). Census **live REST pulse**: SemIf **1984★**; + jevlike **1002★**; TypeAR **11★** flat. Tracker likes + **54** (+3 vs §100 pin **51**); lastModified **CHANGED** + 2026-09-20T02:59:13Z; Hub `archerhume/4rcherhume` HTTP + **401** (not re-fetched as a rewrite). Archer still NOT + landed. Awesomejev 561/27007 user-provided (≠ + AnotiaWang/awesome-jev 84★). + `invented_signal: false`. Do not copy keys / `npm` / + `pip` / `npx` / `uv` / `mix`. No wrapper. No invented + metrics. +- Hourly 1943 HIGH (`research/notes.md` §100): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#17**. Do **not** re-fold 1843 + / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime + / §97 / 1541 / §95 / jev-align *mechanism* / §93. + How-to-apply / mental models / architecture / + Harbor-jevals / toolbelt — not a thin Jev skill dump. + Backend-agnostic categorization/scoring/decision class. + Formal methods compose with scoring; a Noul is a SENSOR; + hard-gating `feels` 80%, pasting 17/18 as Harbor, + hard-gating 0 of 157, treating boolean @ 0.5 as a proof, + claiming 10×, treating softmax as a Noul, or treating 62 + tests as quality is soundness theater. Six HIGH clusters: + **Jev IS the if-statement** + ([southpolesteve/probably](https://github.com/southpolesteve/probably) + PRIMARY; judgments/probabilities drive branches; text + model only writes prose; interpreter owns + variables/loops/budgets/replay; otherwise maybe / + confidence gate; chaos samples after the gate; + southpolesteve/probably ≠ carldaws/hunch ≠ feelings ≠ + Kungie/gut ≠ Illusion47586/judge ≠ tidymodels/probably; + TypeScript MIT; **3★**; HEAD `6bf671a4`; README SHA + `c28570a9`); **GEPA live delta** + ([sutro-sh/jev-align](https://github.com/sutro-sh/jev-align) + 133★ / forks 10 live; build calibrated classifiers from + human feedback; HEAD `49753df9` / README SHA `363fccb7` + unchanged vs §93); **retrieve by relevance not + resemblance** + ([samdotmak/jev-recall](https://github.com/samdotmak/jev-recall) + one calibrated yes/no per memory in one request; pointer + mode 17/18 19/20 *theirs*; embedding resemblance misses + the allergy; MIT; **6★**; HEAD `d3e4acfb`; README SHA + `ea774519`); **memory leases ended by new evidence** + ([chopratejas/invalidate](https://github.com/chopratejas/invalidate) + HIGH upgrade; six Nouls then fixed rules in code; 0 of + 157 false invalidations; questions/plans/directives are + not evidence; unsure → review queue; host keeps the + store; Apache-2.0; **11★**; HEAD `d6ade601`; README SHA + `8cccad5f`); **contract-of-artifact lint rename** + ([mizchi/jev-lint](https://github.com/mizchi/jev-lint) + is mizchi/jevlint rename; name↔body / comment truth / + test-claims; no shipped rule has severity error; ~1 in 5 + findings wrong *theirs*; TypeScript MIT; **13★**; HEAD + `62d73f8e`; README SHA `4c0e37cd`); **JSON Schema + question compiler** + ([Kiln-AI/jev_jsonschema](https://github.com/Kiln-AI/jev_jsonschema) + HIGH upgrade; JSON Schema → typed JSON via Jev; + noul_threshold 0.5 decoder not a proof; + IncompatibleSchemaError lists every bad property; MIT; + **5★**; HEAD `fccea8c2` unchanged); **local System One + economics** + ([mizorewww/laya-coreml](https://github.com/mizorewww/laya-coreml) + on-device Laya CoreML ANE; ~5 ms P50 short decisions; + 189/189 FP16 checkpoint parity; 10× not achieved; + Apache-2.0; **0★**; HEAD `47f4baf0`; README SHA + `2068c661`) + + ([Micha0827/snapjudge](https://github.com/Micha0827/snapjudge) + softmax over allowed tokens ≠ Noul; question-first + cache; Python MIT; **3★**; HEAD `2df5ce27`; README SHA + `64a91458`) + + ([direwolfiy/JevPi](https://github.com/direwolfiy/JevPi) + Jev-first Pi agent loop; slow-LLM fallback; explicit + action menu / CandidateSource unimplemented; 62 tests + wiring not quality; license null; **0★**; HEAD + `980f8895`; README SHA `88ec1a55`). Soft Noul ≠ hard + safety. 0★ HIGH still got a real card. Census **live + REST pulse**: SemIf **1954★**; jevlike **989★**; TypeAR + **11★** flat. Tracker likes **51** flat; lastModified + UNCHANGED 2026-09-19T18:37:18Z; Hub + `archerhume/4rcherhume` HTTP **401** (not re-fetched as + a rewrite). Archer still NOT landed. Awesomejev + 561/27007 user-provided (≠ AnotiaWang/awesome-jev 83★). + `invented_signal: false`. Do not copy keys / `bun` / + `pip` / `npx` / `uv` / HF download. No wrapper. No + invented metrics. +- Hourly 1843 HIGH (`research/notes.md` §99): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#16**. Do **not** re-fold 1740 + / §98 / 1639 / §96 / gliner-native-runtime / §97 / + 1541 / §95. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; hard-gating + cost-derived 0.038 as a proof, merging von Needle + 52.6% with n=78 93%, “guaranteeing” calibration, or + pasting “Jev wins guardrailing” is soundness theater. + Five HIGH clusters: **cost-derived YES/NO/UNSURE + control flow** + ([Kungie/gut](https://github.com/Kungie/gut) + PRIMARY; cost-sensitive decision theory × System One + probabilities → control flow; thresholds derived from + costs not hard-coded; YES / NO / UNSURE from + cost_false_yes / cost_false_no / cost_human; + auto-batching same-object questions; Kungie/gut ≠ + tpellet/hunch ≠ carldaws/hunch; GitHub Apache-2.0 / + LICENSE MIT / pyproject Apache-2.0 *theirs*; **0★**; + HEAD `cb56c875`; README SHA `630474f6`; pre-alpha); + **typed-callback twin** + ([Illusion47586/judge](https://github.com/Illusion47586/judge) + judgment vs generation; deterministic execution after + probabilistic judgment; exactly one app-owned + callback; explicit uncertain branch; + Illusion47586/judge ≠ lexingtonhibiki/judgekit ≠ + Ascurse/typed-judge-kit; TypeScript MIT; **0★**; HEAD + `e69f65a1`; README SHA `08554c6f`; `@brkn-labs/judge` + 0.1.0); **variable-N option scoring as the trainable + object** + ([zwliJay/jev-forge](https://github.com/zwliJay/jev-forge) + dynamic candidate bags not fixed label sets; + zwliJay/jev-forge ≠ NanoJev; class-architecture note, + not a sixth species; GitHub NOASSERTION / LICENSE MIT; + **1★**; HEAD `eb3e4a2d`; README SHA `359f3f57`; do not + clone Hub weights); **open NAR replica economics** + ([wfzyx/von](https://github.com/wfzyx/von) late-catch + HIGH of the same repo; NAR local drop-in; open replica + economics / latency vs closed Jev; competing NAR + claims / replica honesty; Apache-2.0; **43★**; HEAD + `b9e42b26`; README SHA `574aa628`; do not merge Needle + 52.6% with n=78 93.0%); **typed vs chat judges on + guardrailing** + ([ishaannk/llm-vs-jev](https://github.com/ishaannk/llm-vs-jev) + cross-note only; nothing wins outright; can be argued + out of guarding; deeper integrity fold is rh-guard; + Apache-2.0; **0★**; HEAD `182e0864`; README SHA + `d9ebd40f`). gut/judge are **control-flow / + decision-theory overlays, not new class-table + species**. Soft Noul ≠ hard safety. 0★ HIGH still got + a real card. Census **live REST quoted**: SemIf + **1936★**; jevlike **983★** (watch claimed 984); + TypeAR **11★** flat. Tracker likes **51** flat; + lastModified UNCHANGED 2026-09-19T18:37:18Z; Hub + `archerhume/4rcherhume` HTTP **401**. Archer still NOT + landed. Awesomejev 561/27007 user-provided (≠ + AnotiaWang/awesome-jev 83★). X MCP `since_id` held; + pages_archived 0; no invented tweets. + `invented_signal: false`. Do not copy keys / `pip` / + `npm` / `uv` / Hub download. No wrapper. No invented + metrics. +- Effect-oriented loops (`notes.md` §28, `mappings.md` §19): Ward's + ZIO client keeps Jev as the outer Choice and the handler as the + effect. Not Effect.ts. GLiNER author: GLiNER2 "like jev" is GLiGuard + schema-conditioned categorize, not a Noul. +- Boundary-audit stop conditions for TOCTOU-of-Noul and vacuous specs; + FAQ rows for Alloy vs Apalache and PufferLib-as-DST-trio +- Research pointer to [dayhaysoos/jevals](https://github.com/dayhaysoos/jevals): + local MIT workbench for Jev questions vs labeled Noul/Choice/Score cases + (compare runs, WebMCP + agent skill). Empirical acceptance-test surface + for Hypothesis mapping cards; complements `evaluate_decisions.py`. Not a + jevals how-to (`research/notes.md` §24; one sentence in `validation.md`) +- Mental-models card: Augustus is design judgment across AI, SWE, + business, knowledge work, and life — not SWE-only. Pillars: expected + utility / selective classification, calibration and cost-sensitive + thresholds, VOI, MCDA, search/control substitutions, signal detection, + Leveson org/safety, NATM/snap-fit/Norman/Kent/Shirky as general + intuition. Domain gallery labeled Hypothesis except launch-week + Empirical SWE rows. +- Archer Hume architecture reconstruction (17 Sep 2026 essay, ~10k + probes of `jev-1.13.0`): direct readout vs generated confidence, + isolated questions, listwise IIA and order sensitivity, confidence as + arithmetic on the distribution. Independent envelope probe; does not + override live TypeSafe docs. Announced open-weight drop is **WATCH** + (27B dense, AU healthcare residency, prefers "decision models"; still + no Hub weights). `research/notes.md` §31, §33; `judgment-class.md` + when-to-use table; FAQ confidence / surfaces questions. +- Entropy as allocator (**Hypothesis**, `judgment-class.md`): Atallah's + low / medium / high buckets place System One on typed decisions and a + frontier decoder on high-entropy synthesis — same axis as marginals + vs joint and as VOI. "Review this PR" as medium is still partly + generative; "first model ever" is a claim. `research/notes.md` §38 +- Marginals, not a probabilistic program (`judgment-class.md`, FAQ): + Erik Meijer — Jev is a cool API and not a PPL; Kleisli qualifications + exaggerate; "Jev gives you the marginals; a decoder gives you the + joint." Joints and invariants stay with TLA+ / Alloy / contracts. + `research/notes.md` §34 +- Bespoke Nimble: open contrastive recipe, not a Jev distill. Model + card Apache-2.0 LoRA on Qwen3.5-9B (repo license absent). Their + 324-example holdout is a named receipt (Nimble 90.12%, Jev 1.13.0 + 93.21%), not a ranking. 9B-vs-Jev on your labels stays Hypothesis. + `research/notes.md` §35; one sentence in `validation.md` +- djev-spark: third compute graph (diffusion structured reads, + Jev-shaped I/O, image-in). Empirical as the public interface; + Hypothesis that it beats a decision head on your task. Archer's + multimodal drop stays WATCH. `research/notes.md` §36 +- Perception specialist then judgment specialist vs shared multimodal + System One (**Hypothesis**): SAM 3.1 (masks and tracks) or an ASR + transcript, then typed decisions on that state, is an application + pattern, not native omni. Information dies at the interface. Prefer + a shared multimodal decision model when the joint matters (Archer + Watch, not Empirical; djev-spark images; future audio). Basit ask, + primary post not retrieved. `research/notes.md` §39 +- Perception→decision pipeline, measure, and hill-climb + (**Hypothesis**, `validation.md`): stages with a versioned state + contract; stage metrics plus a frozen taskset; HoH changes one stage + or one interface. DSPy/Ax only on LM-program knobs; jevals and + calibration for the decision slice; Harbor names product + end-to-end, not a tutorial. `research/notes.md` §41 +- Eval & hill-climb (`validation.md`): jevals decision-stage hygiene + (independent keys, correctness is not confidence, held-out, immutable + runs) and Harbor as the product taskset substrate; one composition + table. `research/notes.md` §40 + +### Changed + +- Skill description rewritten as trigger conditions (mixed architecture, + prefilter, routing, preference lint, classification skepticism, family + choice including GLiNER/GLiClass/listwise/vision) plus an explicit `not_for` + against the official `typesafe-ai` skill +- Identity lock vs neighbor skills (`typesafe-ai`, `tenbin`, `decision-first`) + so Augustus stays the design-judgment layer — class-wide, not TypeSafe-only +- Design cards name hole, family, and typed judgment provider (Jev default; + other family only with self-eval) +- Protocol fan-out step is family-aware (Jev batch, GLiClass one-pass, + dual-encoder prompt scoring); ranking vs decision fail policy is a + non-negotiable +- Protocol and FAQ branch for "formally verify with Jev"; methods-catalog + and composition-algebra verifier position point at the ownership split +- Skill mission and description are domain-general (AI / SWE / business / + knowledge work / life); FAQ "is this only for software?"; mappings.md + beyond-SWE examples labeled Hypothesis; boundary-audit red flags for + TOCTOU-of-Noul and vacuous specs; formal-methods expanded with Alloy vs + Apalache and the DST trio including PufferLib; GLiNER promoted from + cousin footnote to species-map peer + +### Fixed + +Adversarial review of the whole skill against its own non-negotiables +(findings in `research/notes.md` §27). + +- Gate fail policy is per action, not universally open + (`composition-algebra.md` position 3, `agent-self-assessment.md`): + advisory guards fail open *because* an interlock sits underneath; + selection and authorization gates fail closed +- Dual-orchestration topology A selects from a closed catalog instead of + "planning" MCP calls, which contradicted the standing planner rejection +- Species map applied to the skill's own advice: GLiClass (categorize) is + the large-catalog substitute for a 255-option Choice; GLiNER spans are + not (`SKILL.md`, `judgment-class.md`, `applied-mappings.md`) +- Han Xiao trolley relabeled an Empirical **rejection** (one tweet, no + repo), not a recipe +- openjev-lm caveat moved to the figure it belongs to: 92.9% is against 70 + hand-labelled gold, 98.1% is teacher *agreement* +- Contract surface removed from design cards: the Ax constructor call and + the `instructions` key enumeration point at live docs instead + (`optimizer-integration.md`, `question-design.md`) +- `mappings.md` preamble no longer claims uniform Hypothesis where card + bodies say Contract/Empirical; §17 forbids reusing jevgate's ≤0.18 as a + constant; all Hypothesis-range references aligned to §6–§19 +- Ownership split labeled Contract in `toolbox-mapping.md`, matching + `mappings.md` §8; done-check splits structure from the Noul + +Second pass on `7b3a0c3` (`research/notes.md` §43). Zero blockers. +Dropped the unpublished `npx jevals` line; SAM and ASR are upstream +producers, not the perceive species; removed two call shapes from +`optimizer-integration.md`; tagged the $0.042/MTok cell as a vendor +figure; marked GodsBoy 94.4% exploratory. +- Skill description gained trigger terms for boundary audit, question + diagnosis, agent self-supervision, and optimizer placement + +## [0.2.0] - 2026-09-18 + +Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) +(`65a39f3`, 2026-09-12). + +### Added + +- Boundary-audit card for existing systems: three-way split (exact / + bounded judgment / generation), code-smell catalog, fit test, opportunity + map, smallest-viable-boundary rule, Jev-around-LLM sandwich, centralized + policy + raw-judgment retention, red flags, completion questions +- Protocol branch: audit a codebase/PR before inventing mappings; per-action + risk gates; keep questions/thresholds in one reviewable module +- Skill description trigger terms for brittle parsers, prompt-to-JSON + classifiers, and agent loops that are really bounded decisions + +## [0.1.0] - 2026-09-18 + +Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) +(`65a39f3`, 2026-09-12) — the only tagged revision of the official skill at +Augustus launch. + +### Added + +- Skill protocol, decision-design card, and evidence labels (Contract / + Empirical recipe / Hypothesis) +- Classical-method mappings: features/utility, selective decisions, decision + circuits, bounded rerank, hierarchy/beam search +- Agent self-assessment, optimizer coupling (Ax, DSPy, ProgramAsWeights) +- Toolbox sweep, named-methods catalog, 11-position composition algebra, + question-design diagnosis +- Validation gates and `scripts/evaluate_decisions.py` +- Launch-week evidence archive (187 repos) and public ecosystem index +- Claude Code marketplace manifest; public GitHub mirror at + [`24601/Augustus`](https://github.com/24601/Augustus) + +### Research log (pre-tag) + +The dated passes below are how 0.1.0 was assembled. + +### 2026-09-18 (refresh pass 2) +- Research: Gemini Deep Research report retrieved and archived (interaction ID + saved in jev-archive/state); GitHub census doubled to ~60 Jev repos + + framework integrations (LiteLLM, LangChain, Vercel AI, Mastra, Eliza, Ax, + Composio); all 87 repos cloned to /home/user/workspace/jev-archive for + hourly refresh. +- New measured recipes added to notes.md: foreman supervision loop, pi-jev + gate thresholds, pi-warden 6→0 paired-run result, winnow relevance sieve, + fast-jev-compaction two-noul rule, skill-router gates (0.30/0.40, shortlist + 3, 94.4% vs 70.8%), calibration ECE 0.0313 vs 32% OOD collapse (Archer + Hume), Every 777-judgment eval, Near Here moderation numbers. +- Skill: added references/agent-self-assessment.md (agent self-supervision + lifecycle, grounding/citation checks, skill callability testing) and two + mapping-index rows; validation.md dogfooding section still canonical. + +### 2026-09-17/18 (initial) +- Baseline research archive (sources.json, notes.md), augustus skill with + mappings + validation references, evaluator script, hourly refresh script, + Claude plugin marketplace manifest. + +### 2026-09-18 (topic-index pass 3) +- Fixed census method: exact GitHub search paginated (700 repos created since + 09-14 captured; 700-result cap noted) + topics/jev crawl → ~80 additional + repos; archive now 184 clones. Miss-cause documented: earlier star-sorted + limit-40 search cut the low-star tail (incl. both MCTS repos). +- Skill: MCTS mapping promoted experimental → empirical recipe (grounded vs + speculative fidelity in types; probes-only concession; measured 24/24 vs + 1/24 greedy); agent-self-assessment.md gains the judge-variance recipe + (Jev judge 224-279x more consistent than LLM judge over 100 reps). + +### 2026-09-18 (pass 4 — optimizers + official skills + clone audit) +- ax Jev support documented from source (native adapter details, trueThreshold + semantics, fail-closed mapping validation); new reference + optimizer-integration.md covering Ax + DSPy typesafeify + jev-dspy-lab. +- typesafeainate/dspy-typesafeify cloned; official typesafe-ai/skills already + archived and layered-on (never duplicated). +- Clone audit: repos.txt deduped (185 unique), 0 missing on disk, no failures. + +### 2026-09-18 (pass 5 — toolbox sweep meta-method) +- New references/toolbox-mapping.md: the how-to-find-approaches-and- + applications procedure (judgment-shaped-hole substitution, newly-feasible + classification via economics inversion, standing rejections list); wired + into SKILL.md central model + index row. + +### 2026-09-18 (pass 6 — named-methods + operators/theorems tier) +- references/methods-catalog.md: ~20 named algorithms (CatBoost row is + Empirical via autoresearch cookbook) + operators/theorems tier with + precondition-carrying rule; wired into SKILL.md index and toolbox sweep. + +### 2026-09-18 (pass 7 — composition algebra as application generator) +- references/composition-algebra.md: 11-position grammar of Jev-vs-construct + relations, logical-operator combination rules, and the position×construct + traversal as the systematic application generator; wired into SKILL.md + index + toolbox sweep. +- Hourly 2340 HIGH (`research/notes.md` §104): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#21**. Do **not** re-fold 2246 + / §103 / 2145 / §102 / 2041 / §101 / 1943 / §100 / 1843 + / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime + / §97 / 1541 / §95 / jev-align *mechanism* / §93 / + jev-orderby-bench *six-gates* / §60. How-to-apply / + mental models / architecture / Harbor-jevals / toolbelt + — not a thin Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating a + description-only stub as a checkpoint, re-folding + six-gates as new, distilling Jev as teacher of record, + pasting openjev-sglang as OpenJevPro, treating + constrained logprob as a Noul, quoting an + untrained-looking demo as Jev identity, pasting a + radar's listed numbers, or quoting 0.916 as a class + ceiling is soundness theater. Distill-Jev UI stub and + constrained-logprob-as-Noul are the anti-patterns. + Eleven HIGH clusters: **from-scratch calibrated + decision model** + ([hyusi2003/MiniSystemOne](https://github.com/hyusi2003/MiniSystemOne) + PRIMARY; train calibrated ~27M from scratch; typed Q→prob dist / one forward pass / no LLM decode; + hyusi2003/MiniSystemOne ≠ Colvin0315/MiniSystemOne; description-only stub / size 5; + Apache-2.0; **0★**; HEAD `4385335b`; README SHA `83016bf8`); + **ORDER BY ranking upgrade** + ([yodablocks/jev-orderby-bench](https://github.com/yodablocks/jev-orderby-bench) + ESCI hard probe fails four of six; jev_bool ECE 0.242 inversion 0.255; + do not re-fold §60 six-gates as new; jobbyjev one-request-per-company from batch-size result; + Python MIT; **0★**; HEAD `52397954`; + README SHA `7bd075c3`; size **309**); + **find/design/evaluate decision loops** + ([karanb192/jev-architect](https://github.com/karanb192/jev-architect) + find/design/evaluate TypeSafe Jev decision loops; karanb192/jev-architect ≠ samtay32/jev-system-architect; + HTML MIT; **0★**; HEAD `35ea6d93`; README SHA `68c2b5f9`; size **5199**); + **distill-Jev UI stub** + ([Jairik/jev-distiller](https://github.com/Jairik/jev-distiller) + Jairik/jev-distiller size 1; distill-Jev UI stub / do not distill Jev as teacher of record; + MIT; **0★**; HEAD `0589d44c`; README SHA `aa408c5e`); + **post-launch scored opportunity map** + ([licensedsaucer9-web/jev-opportunities](https://github.com/licensedsaucer9-web/jev-opportunities) + post-launch scored use-case map / Jev self-scores then human curation; + license null; **0★**; HEAD `a47fa414`; README SHA `aa33f901`; size **27**); + **Jev-inize a use case** + ([gavinHuang/jevinize](https://github.com/gavinHuang/jevinize) + Jev-inize a use case into classifier/router; gavinHuang/jevinize → simple-jev not TypeSafe; + featherless-ai/simple-jev; + MIT; **0★**; HEAD `6d080632`; README SHA `5f48e622`; size **6**); + **saved-decision regression** + ([VihaanAgarwal/jev-diff](https://github.com/VihaanAgarwal/jev-diff) + compare saved decisions / same label can still change the branch; + VihaanAgarwal/jev-diff ≠ Saik0s/diffusiongemma-jev-macos; + not tested with a live Jev API key; Python MIT; **0★**; HEAD `a3c98807`; + README SHA `3b0ce75c`; size **120**); + **constrained-logprob API** + ([zhangcy122/OpenJevPro](https://github.com/zhangcy122/OpenJevPro) + constrained logprob + temp/Platt ≠ Noul; OpenJevPro pastes openjev-sglang JevBench as own; + zhangcy122/OpenJevPro ≠ IamBusy/OpenJev ≠ ekzhang/openjev-sglang; + PolyForm Noncommercial; HTML; SPDX NOASSERTION; **0★**; HEAD `94d77bcb`; + README SHA `50c77ace`; size **64** (relock; was **62**)); + **SmolLM RLCD reproduction** + ([patelvishwa112/jev-system-one-rlcd](https://github.com/patelvishwa112/jev-system-one-rlcd) + SmolLM-135M / sub-70ms / 0 output tokens; demo P(True) 0.5052 / Choice conf 0.2872 / Score conf 0.0055; + README claims MIT / GitHub license null / no LICENSE file; + patelvishwa112/jev-system-one-rlcd ≠ arnabgho/rlcd-lite ≠ blackwood-rlcd; + Python; **0★**; + HEAD `62b103b3`; README SHA `55994d69`); + **source-backed Awesome radar** + ([logicrw/awesome-jev-projects](https://github.com/logicrw/awesome-jev-projects) + source-backed Awesome Jev radar / 306+ commit-pinned; auto GitHub sync / Issue-only submissions; + logicrw/awesome-jev-projects ≠ AnotiaWang/awesome-jev ≠ yibie/awesome-jev ≠ cobanov/awesome-jev ≠ rupeshpoojary9/awesome-open-system-one; + JavaScript MIT; **136★**; HEAD `97057cc1`; README SHA `25a19b31`; size **7677** (relock; was **7136**)); + **rival-aware one-pass scorer** + ([olanotolu/jevbetter](https://github.com/olanotolu/jevbetter) + hashed n-gram encoder / rival-aware attention; olanotolu/jevbetter vs jevlike starter; + synthetic hard menus top-1 0.916 vs 0.873 / ECE 0.0182 vs 0.0367 / 40 vs 4608 menus/sec; + shuffled-context control 0.335; Python MIT; **12★**; HEAD `bb0ebc82`; + README SHA `5cbe01d4`; size **324**). + Soft Noul ≠ hard safety. 0★ HIGH still got a real card + (awesome-jev-projects **136★**; jevbetter **12★**). Census + **live REST pulse**: SemIf **2047★** (+28 vs §103 **2019**); + jevlike **1018★** (+12 vs §103 **1006**); TypeAR **12★** + **flat**. Tracker `multimodalart/jev-reproductions-tracker` + likes **56**; lastModified + `2026-09-20T04:29:16.000Z`; Hub `archerhume/4rcherhume` + HTTP **401** (not re-fetched as a rewrite). Archer + still NOT landed. Awesomejev 561/27007 user-provided + (≠ AnotiaWang/awesome-jev **91★** ≠ logicrw **136★**). + Qwen3.8-27B ≠ Archer (likes **15787**). + Independent adversarial review relock (PR #22 + after `babb111`): HIGH HEAD/README unchanged; + star/size lag locked (awesome **136★**; SemIf + **2047★**; jevlike **1018★**; AnotiaWang **91★**; + OpenJevPro size **64**; simple-jev **318★**; + openjev-sglang **205★**; RLCD size **1513**; + Colvin size **814**; yibie **430★**; cobanov **224★**). + `invented_signal: false`. Do not + copy keys / `npm` / `pip` / `npx` / `uv` / + `.env`. No wrapper. No invented metrics. +- Mixed-architecture card: default placement is judgment-class model + + generator + code, not stack replacement. Covers cost-sensitive prefilter + (fail-open vs fail-closed per action), tool/skill routing, AGENTS.md + preference lint, a placement gallery from the 2026-09-18 X+GH hour, and + an explicit answer to "Jev is just classification" +- Protocol branch and mapping-index rows for those four placements +- Non-negotiable: classification is not the product; typed judgment is a + software primitive placed beside generation +- Hourly research archive for this pass (X theme digest + `topic:jev` movers) + under `research/archive/hourly/2026-09-18T14/` +- Research note on Laya (`convaiinnovations/laya`): open Choice/Score/Noul + head as a self-hosted *typed judgment provider*; vendor benches labeled + claims; TypeSafe remains the default path +- Applied-mapping cards: context sieve, exact-text keep/drop, env/harness + triage, moderation/ranking, skill/tool routing (`applied-mappings.md`) +- FAQ card for "it's just classification", stack replacement, Jev vs open + head, and Augustus vs neighbor how-to skills +- Judgment-class card: Augustus covers the whole class of fast/cheap + categorization-classification-scoring models (Jev is exemplar, not + monopoly). Families: closed decision API, open System-1 heads (Laya), + GLiNER/GLiClass encoder family (locate vs categorize vs local + multi-head), listwise/pairwise rankers, vision scorers. + Species map in `judgment-class.md`; GLiNER is a peer, not a footnote. + Fork of listwise discriminative vs decision/proper-scoring objectives; + four vision scoring patterns; seven portents for agent architecture. + FAQ rows for family choice, GLiNER vs GLiClass vs Jev vs cross-encoder, and + CLIP/SigLIP gating. No invented APIs. +- Formal-methods card: judgment vs proof ownership (sensor / constraint / + searchlight); Alloy Analyzer vs Apalache (model finder ≠ SMT BMC ≠ + inductiveness); TLA+/Quint/P/NuSMV/PRISM/Event-B; Dafny/JML/ + Frama-C/SPARK; DST trio (Antithesis hypervisor, Resonate Lean+oracle+ + SDK, PufferLib env+seed / Ocean trainer contracts); harms + (TOCTOU-of-Noul, soundness theater, AI×FM / Hillel vibing specs); + crossover metaphors (NATM, snap-fit, Norman gulfs, Leveson STAMP/STPA). + Curriculum archived at `research/archive/curriculum/FORMAL-METHODS-SYSTEM-ONE.md`; + named rows folded here. One-screen alias: `formal-semi-formal.md`. + Non-negotiable: never launder a Noul as a proof. +- One-screen `references/formal-semi-formal.md` (curriculum 1-pager) +- Hypothesis mapping cards (do not promote without an acceptance test): + VOI / gather; SDT/ROC; Leveson sensor≠constraint; search/control + outside SWE; spec property pipeline; Alloy instance loop; runtime + assurance sandwich; DST multiverse triage; durable agent control; + assignment hybrid; situated density (`mappings.md` §6–§16) +- Input-brittleness and structural-prove ∩ remainder cards + (`mappings.md` §17–§18): paraphrase pairs → Chow abstain; allowlist / + text-layer first, judge leftovers (jevgate / doc-router *shapes* + Empirical; domain-general reading Hypothesis). FAQ: GLiNER vs Jev, + LLM-as-judge (Langfuse framing), allowlist-then-judge +- GLiGuard as an Empirical encoder peer (`judgment-class.md`, + `notes.md` §30): one-pass safety-schema classify on GLiNER2, not a + Jev weight clone; FAQ "is GLiGuard Jev?"; README OR/refusal + aggregation left as existing policy-in-code. LLM I/O safety is not + a coding-agent tool gate +- Hourly 10:07 Boise fold (`research/notes.md` §25–§26): GLiNER2.5 local + peer; openjev-lm 92.9% / 6 vCPU teacher-distill; jevgate; doc-router + 1.74× $; pi-jev-context; jevscope next to jevals; Han Xiao trolley + (listwise ≠ decide); James Ward dual orchestration; JevLint +- Constrained-AR surface, not a sixth species (`judgment-class.md`): + TypeAR puts a typed interface on a pretrained generator (next-token + constraint ≠ proper-scoring head). Archer Hume's open-weight drop + stays **Watch** (`research/notes.md` §31, §32) +- Hourly ~11:02 Boise fold (`research/notes.md` §33): Archer + clarifications still Watch (27B dense one-forward-pass, multimodal + generalization report, AU healthcare residency not anti-TypeSafe, + prefers "decision models"); when-to-use table (proprietary Jev vs + Archer vs TypeAR vs encoder DeBERTa vs LoRA distill); HF novel + (jev-gate-student-b 148k corpus, jp-sns-jev7 ONNX, open-jev-deberta, + mini-jev-runs 27.9k logits, jev-tree-choice-cap); device/harness + (jev-mobile MCP, jev-macos-loop, jev-harness, routeKit); HacksonClark + SREGym-Lite 20/50→24/50 — rank tests, do not diagnose +- Hourly ~11:59 Boise fold (`research/notes.md` §42): Archer still + Watch. Three open paths (encoder / AR constrained decode / trained + decision-only). Native constrained serving + ([pcdServer](https://github.com/stephanj/pcdServer), TypeAR-class, + 2–256 enums, Apple+Linux GGUF). Meta-VOI hook + (typesafe-jev-tools 149-row: Haiku more accurate, Jev confidence + monotonic). jev-mode latency-class split (token ratio durable; + accuracy is parity). OpenSmoke env-break vs policy-break + + pre-mortem. jevql store-as-decision-surface. jot topology B with a + closed catalog. openevals online full-traffic. hermes north-star + two-layer finish gate. pi-jev (not pi-jev-context). jev-plays-games + option-order probe. joxide jump-by-description. laya-typed-decisions + companion packaging. No wrapper. +- Hourly ~12:58 Boise fold (`research/notes.md` §44): Archer still + Watch (no architecture rewrite). Store-index fork: in-engine + ([sqlite-jev](https://github.com/mgaitan/sqlite-jev), pg-jev cousin) + vs CLI rewrite (jevql). Soft judgment inside a hard envelope + ([bitrate-advisor](https://github.com/affirmitv/bitrate-advisor); + mmalisper JOB planner +12% geomean, author-reported; join-order + Choice alone was 2× slower). Distill-to-device as a *memory* gate + (jev-gate, already §33). Encoder vs decoder open-replica receipts + (openjev-lm $0/call overnight CPU). jev-harness as Harbor-adjacent + practice (assert on action). Host adapter + ([jev-routing](https://github.com/nekowasabi/jev-routing), not MCP); + OpenClaw typed routing ([jev-claw](https://github.com/trietphan/jev-claw)). + Voice-control and JevML are README stubs. Higgsfield auto-routing is + a claim. No wrapper. +- kev (`jaredpalmer/kev`, `research/notes.md` §45): runnable Archer + reconstruction on the trained decision-only open path next to Laya / + Nimble / Watch. Qwen2.5-0.5B LoRA + pointer, Apache-2.0, `POST + /v1/systemone` drop-in. Isolation exact (packed vs separate max Δ + 3.7e-6; secret-in-sibling p=0.03 vs in-state 0.99). Held-out ECE + 0.065 (0.031 after temp scale); acc 0.799 on 1,350 ID questions. + Permute argmax flips 7.4%; IIA log-odds shift mean 0.13; boundary + forgery held. Laptop-local System One for development/eval; not a + knowledge/frontier substitute; not a Jev teacher-copy. Contrast vs + TypeAR, encoder DeBERTa, proprietary Jev. jevals/Harbor bake-off + candidate. No serve how-to. +- Hourly ~14:03 Boise fold (`research/notes.md` §46): Archer still + Watch. Open multimodal RLCD + ([blackwood-rlcd](https://huggingface.co/BlackwoodAI/blackwood-rlcd), + CC BY-NC): screenshot + marked candidates → Choice; web acc 0.907 vs + Jev 1.13 text-only 0.480; letter-shuffle 0.133 vs 0.587; ECE 0.037; + ~200 ms H100; Jev still leads general text 0.850 vs 0.786. Shared + bake-off ([open-jev-laya-bench](https://huggingface.co/datasets/pngwn/open-jev-laya-bench)): + 26+9 tasks, 11959 items; ECE/NLL/Brier; macro acc Δ +0.023 + neutral / +0.229 home; LLM-as-judge is not the score. Decision-token + QLoRA + ([Foodoo1/Qwen3-14B-RLCD-Decision-LoRA](https://huggingface.co/Foodoo1/Qwen3-14B-RLCD-Decision-LoRA)): + fraud_risk 64→95%, overall 85.2→98.8% at ~234 ms/4-field broadcast; + synthetic. jevgate frame: allowlist *proves*, Jev judges only + unlisted, fail-open. wellposed: missing `other` → confidence 1.00 + wrong; gating cannot catch it (`tenbin` owns the lint skill). + S1 reflex keeps control (jev-reflex-autonomy-lab). MED: + jev-decision-layer, jev-e2e, jevpandas. No wrapper. +- Abide (`coldteadotai/abide`, `research/notes.md` §47): productized + Jev preference lint for Claude Code / Codex / OpenCode. Soft + AGENTS.md / CLAUDE.md rules → one Score per rule on the diff (never + the conversation); hard rules stay with the linter (same layering + family as jevgate). Edit- vs turn-phase observation window; banded + confidence (≥0.8 repair / 0.5–0.8 note / <0.5 silence — their + operating point) + fail-open hooks; rubric.json quotes source + lines; calibrate/tune fix false positives in the question. Replay + of 93 sessions (1,256 edits / 147 turns) with independent review: + edit precision ~26%, turn ~73% (author-reported, before tune). + Fuller productized path of the jev-pref contract. Complementary to + rh-guard (eval-integrity vs project soft rules). Text/diff only — + not multimodal. No hook how-to. +- kev delta (`research/notes.md` §45): Hub weights + [`jaredpalmer/kev-0.5b`](https://huggingface.co/jaredpalmer/kev-0.5b); + `--run` accepts Hub ids; PEFT `task_type=FEATURE_EXTRACTION` (publish + patches legacy adapters). HIGH question-design: confront Choice + `"other"` / none-of-the-above as a wrong alternative too, vary + wording, dedicated `none_of_the_above` eval (no published rates). + Cross-link wellposed request-shape lint. No species change. No + wrapper. +- Hourly ~14:52 Boise fold (`research/notes.md` §48): Archer still + Watch. Extractive selection + offline `redecide` + ([testimonial-miner](https://github.com/AppitStudio/testimonial-miner)); + pointer-not-generator + ([jev-reviewer](https://github.com/choxos/jev-reviewer)). Local + `/v1/systemone` drop-in ([jev-local](https://github.com/us/jev-local); + default scorer is a stub until `hf`). Observe→decide→verified-act, + no screenshots ([solari-reflex](https://github.com/hitakshiA/solari-reflex); + 60.2/194.9, 66/460, 24.2/98.4 s vs Codex on Solari). Dataframe + accessor sibling ([jevframe](https://github.com/ktaletsk/jevframe); + note jevpandas). Route ≠ memory (jev-hermes). Advisory sidecar + (agent-workflow-typesafe-ai). Structure induction (dag-jev experiment). + Decision-for-control / generator-for-content (jev-agentworld-web-simulator). + Collab arms + Wilson/McNemar (jev-testbench). AST ∩ semantic (jevscan; + `tenbin` owns lint). Light Pi gate (pi-jev-approver). Laya ONNX port + ([laya-onnx](https://huggingface.co/Mattepiu/laya-onnx); do not copy + vs-Jev table). Spotcheck: SemIf 1551★; jevlike 866★; tracker + 20:12:57Z still lists Laya, not Blackwood. No wrapper. +- Hourly ~15:52 Boise fold (`research/notes.md` §49): Archer still + Watch. X discourse blocked. Boundary map / extractable-from-state + ([jev-capability-atlas](https://github.com/Zaious/jev-capability-atlas); + history suite A wrong@0.90 / B 0.07 / C right@0.97; component node; + dangerous-high ECE; DOM-as-text + fan-out). Harbor-style bake-off vs + constrained LLMs + ([DMB](https://github.com/nibzard/decision-model-benchmark) v2: jev + banking 76.3% / spam 93.0% / 256+ cap; p50 264–276 ms; $0.07/1k; no + class wins on quality). Feedstock + ([jevals-data](https://github.com/Jevals/jevals-data) CC-BY-4.0; + recompute-from-logs; 2026-09-18 board). Dual-process S1 decide / S2 + generate ([dual-process-ai](https://github.com/taro1985/dual-process-ai); + routing accuracy unmeasured). Combinatorial ≠ extractive (ARC-AGI + Direct Jev 4/400). Packed one-forward open LLM + ([open-alternative-jev](https://github.com/ikermoel/open-alternative-jev) + RACE-H 92.9% @ 4.55 q/s; not a Jev reproduction). Tiny SAN local + surface ([von](https://github.com/wfzyx/von) 14 MB; not a replica). + kev light delta **100★**. Do not merge Banking77 87% / 76.3% / + 79.67%. No wrapper. +- GLiNER2.5 extractive compaction (`research/notes.md` §50, + [gliner25-compaction](https://github.com/m-newhauser/gliner25-compaction), + Apache-2.0): architecture notes, not a plugin how-to. Pointer + keep-drop (character-offset copies) vs generator summarizers; + family with testimonial-miner / jev-reviewer. Soft retention Choice + under a hard mutation envelope (mutating tools / shell operators → + `keep_full`); low-confidence / invalid evidence fail closed to + `keep_full` — contrast many fail-open Jev gates. Same compaction + *job* as fast-jev-compaction / pi-jev-compaction; GLiNER encoder + backend; Fastino/GLiGuard sibling class. `shadowMode` default true. + Not Jev. Not multimodal. No invented metrics. +- CI merge-gate / fail-open wake VOI / S1 indexer / claim-evidence + (`research/notes.md` §51): architecture notes, not a plugin how-to. + [latch](https://github.com/CaseReed/latch) cluster-then-policy + PASS/BLOCK (pair Harbor + rh-guard). + [wakegate](https://github.com/shitianfang/wakegate) skip only if + p(wake)<0.2 (21/21 smoke). s1-graphify-indexer GLiNER extract + + escalate-S2 (10–50× unfilled). + [clear-head](https://github.com/VladyslavHontar/clear-head) + claims vs session evidence. + reification-labs/foreman description-only Phoenix scaffold (not the + super-jev loop). + [jev-gateway-bench](https://github.com/vinilana/jev-gateway-bench) + Harbor on/off one-run signal. + jev-marshal Watch/empty; jevons bounded Pi supervisor (shadow + recovery). MED: if-ai, omp-auto-mode, downloads-sorter, label-desk, + herdr-jev. Archer still Watch. No invented metrics. No wrapper. +- GLiNER2 Ultrafast observe→score→act (`research/notes.md` §52, + [gliner2-ultrafast](https://github.com/sahibzada-allahyar/gliner2-ultrafast), + MIT): architecture notes, not a browser-agent how-to. Same + observe→score-among-candidates→code-acts *job* as jev-ultrafast / + solari-reflex; local GLiNER2 (`fastino/gliner2-multi-v1`) backend, + not GLiNER2.5. No screenshots; no generated selectors; code owns + actuators. Hybrid local decide + remote fill (Mercury 2.5 default + for TYPE). `DONE` ≠ verified success. Contrast blackwood-rlcd + screenshot multimodal; laya-mind2web is DOM-index Laya (same + observed-candidate family). Fastino sibling class with + gliner25-compaction (different hole) and GLiGuard (safety schema). + Demo (theirs, not re-run): Flights 12.20 s / 13.785 s / ~$0.0001 + API — demonstration, not a bake-off. No invented metrics. +- jev-pruner evidence-preserving Bash stdout prune (`research/notes.md` + §53, [jev-pruner](https://github.com/tamaratran/jev-pruner), MIT): + architecture notes, not a plugin how-to. After Bash, Jev Noul-prunes + stdout chunks before the main LLM sees them — no summary. Hard + envelope (≤10k estimated tokens / JSON-diff-whole-doc untouched) + then soft Noul; fail-safe keep original; full archive. Marketplace + id still `fast-jev-output`. Codex is opt-in wrapper, not automatic + interception. Same evidence-preserving *family* as + fast-jev-compaction and gliner25-compaction; different *job* + (command output vs session memory) and Jev backend vs GLiNER2.5. + Manual sweep (theirs): needles 24/24; mean reduction 83% on trim + scenarios. Harbor plugin-eval cannot reach Jev. Terminal-Bench + paired pilot is integration, not a full bench. No invented metrics. +- Cua-S1 specialist System One computer-use (`research/notes.md` §54, + [cua-s1](https://github.com/trycua/cua/tree/main/libs/cua-s1), + parent MIT, ~23.3k★ this pass): architecture notes, not a Driver / + MCP / `uv` how-to. Form-oriented profile `cua-s1-form-v0`. Byte + encoder + option-attention head chooses fill/check/click/skip per + observed element; does not generate values or selectors. Plan ≠ + execute; dry-run default; `execute`/`submit` independent opt-ins; + fail-closed on unknown checkbox / fill without advertised token + `set_value`. **Not TypeSafe Jev** — parallel "System One" naming in + CUA research. Same observe→score-among-candidates→code-acts *job* + as jev-ultrafast / gliner2-ultrafast / solari-reflex / laya-mind2web; + specialist form contract, source-only this pass (no weights, no + checkpoint scores). Offline metric *names* only (accuracy, + abstention, coverage, wrong actions/targets, unsafe when should + abstain). Tests exercise implementation, not checkpoint quality. + Watch for a `cua-s1-form-v0` artifact drop. No invented metrics. +- Hourly ~17:48 Boise fold (`research/notes.md` §55): Archer still + Watch. X MCP flap; `since_id` not advanced. Architecture notes, not + a how-to. Local CUDA/PyTorch Choice/Score/Noul replica + ([jevify](https://github.com/Mintzs/jevify); uncalibrated + likelihoods ≠ Noul; no LICENSE this pass; independent of + Distillation). Decision-native RAG + ([decision-native-rag-skills](https://github.com/emergency-lee/decision-native-rag-skills); + retrieve wide → decide → evidence set; no bundled harness; no + universal benchmark). Verbatim session ledger + scored recall + ([carryforward](https://github.com/Dharundp6/jev-carryforward); + rules never judged; fail-open dump; 9×3 hint). Judgment as a + Ruby language primitive ([hunch](https://github.com/carldaws/hunch); + English-as-config; `rescue nil` fail-open at save). Healthcare + Harbor-shaped S1+S2 + ([explore-typesafe-ai](https://github.com/si618/explore-typesafe-ai); + synthetic FHIR; not clinically validated). Pre-registered + independent eval + ([jev-baselines-eval](https://github.com/ickma2311/jev-baselines-eval); + **both AMBIGUOUS**; cascade sign-flip at exact parity; + confidence=1.0 theater; encoder-with-labels wins; serving-path ≠ + model-speed; same-day errata ×3). Student-b light delta only (HF + card unchanged). MED: toolgate (pre-exec allow/block/review; Jev + not authorization), typesafe-screening-mcp (PubMed screening aid), + databricks-jev-pdf-lab (**honest negative**; no OSS license), + yannip1234/codex-jev (extractive compression family; equal + accuracy/lower cost not established), kazuhideoki/jev-search + (recursive *file* search + fzf; **not** superagents-lab web + search). No wrapper. No invented metrics. +- Hourly ~18:38 Boise 2026-09-18 / 00:38 UTC 2026-09-19 fold + (`research/notes.md` §56): Archer still Watch. Architecture + notes, not a plugin / showcase catalog. Classify-first MCP + ([jev-sift](https://github.com/kbhuw/jev-sift); batch path/url/text + → Jev without entering main agent context first; 50 / 60k / 2MB / + public-IP envelope; mocks ≠ accuracy; no LICENSE this pass; + topology A MCP, not jev-routing). Living applied-mappings atlas + ([jevable.com](https://jevable.com/); claimed 342 vs JSON-LD first + page 36; class patterns — intent columns, score-among-observed, + VOI gates, generative UI decide, robotics text-state, draft-gate + silence ≠ safer — not a 342-title hit list). Maker clocks stay + claims unless already a named receipt. No wrapper. No invented + metrics. +- Stagehand experimental Jev stack (`research/notes.md` §57, + [#2955](https://github.com/browserbase/stagehand/pull/2955) 5/5 of + #2951–#2955, all OPEN draft): architecture notes, not an SDK + how-to. Major harness productization of + observe→score-among-candidates→code-acts (cousins jev-ultrafast / + gliner2-ultrafast / cua-s1 / solari). Jev picks a11y elements; + code copies text. extract `"off"` | `"judge"` | `"pick"`. Their + card (gemini-3.8-flash, 25×3): **37/75** no-LLM ~0.5 s vs baseline + **4.37 s**; 69/75 vs 23/25 (92% both); LLM-off **36/75** — pick is + a fast path, not a replacement. Screenshot extract always LLM. + Cache-check errors never block replay. Do not merge clocks. No + invented metrics. +- Hourly ~18:46 Boise 2026-09-18 / 00:46 UTC 2026-09-19 fold + (`research/notes.md` §58): Archer still Watch. Architecture + notes, not a Convex / uv / pnpm catalog. Public judgment wall + ([ask-jev-ai](https://github.com/waynesutton/ask-jev-ai); 6 + parallel questions; policy-in-code; cost-to-1M from tokens; + license null). Meaning-search without embeddings + ([jevgrep](https://github.com/Bentlybro/jevgrep); 79% top-5 vs + BM25 40% / grep 20% on stripped repos; keyword still wins exact + strings). PR attention ≠ correctness + ([egma-ai/jev-reviewer](https://github.com/egma-ai/jev-reviewer); + **not** choxos pointer-not-generator). Skills→oxlint + ([jev-oxlint](https://github.com/cephalization/jev-oxlint); + AST prove ∩ remainder; Phoenix fixtures; not a hard gate; + `tenbin` owns lint). Session-sticky first-prompt routing + ([jev-adaptive-thinking](https://github.com/jxu-dev-c/jev-adaptive-thinking); + fail-closed fallback). Measured RAG rerank + ([Jev-RAG](https://github.com/Max-sm-yc/Jev-RAG); one-run ≥70% + cost / 72% latency vs Spark *rerank*; full-context Spark still + faster). MED: safe-sh, jev-loan-triage, TurboGuo arenas, jevbox; + hermes/mcp packs not found this pass. No wrapper. No invented + metrics. +- Hourly ~19:48 Boise 2026-09-18 / 01:48 UTC 2026-09-19 fold + (`research/notes.md` §59): Archer still Watch. Architecture + notes, not a pip / venv / Cloudflare catalog. Capability kernel + ([interlock](https://github.com/somoore/interlock); LLM ring 3 / + kernel ring 0; secrets never in the agent; Jev SENSOR; + `policy.py` BLOCK/ASK/ALLOW; type-safe ≠ correct; distinct from + toolgate). Typed control plane around DSPy + ([jev-dspy-control-plane](https://github.com/manikanda-kumar/jev-dspy-control-plane); + DSPy drafts AFTER route+action; OpenJEV / DSPy / JSON Schema + share ontology; offline heuristic ≠ quality). Native-probability + calibration arena + ([jev-arena](https://github.com/meetr1912/jev-arena); live 145 + noul Brier 0.0059 / ECE 0.0620 *theirs*; overconfident in low + bins; 2-request fan-out) plus sonar (heatmap-as-policy) / + vickrey (Jev never bids) / bracket (Brier vs Elo; live trailed + Elo). Engine owns truth / Jev owns judgment + ([game-coach](https://github.com/JoelLewis/game-coach); Wave 0 + PRD; Stockfish WASM; GPL-3.0; anti-soundness-theater with egma). + Human-confirmed port cleanup + ([port-cleanup](https://github.com/epiphany-dynamics/port-cleanup); + Jev recommends; human is the only kill trigger; identity + re-check; shields; mapped explanations). MED toolbelt: + jev-pr-labeler, jevcumber, typedecide, jevon, dsh-jev, + fast-jev-compaction-pi, jev-tetris-benchmark, modelsystem, + opencode-system-one, browser-ai, semantic-bookmark. Skip + jef-mcp (parody) and jevregist (account farming). Star spike: + SemIf 1491→1606 (this pass 1607); jevlike 851→896 (this pass + 897). No wrapper. No invented metrics. +- Hourly ~20:43 Boise 2026-09-18 / 02:43 UTC 2026-09-19 fold + (`research/notes.md` §60): Archer still Watch. Architecture + notes, not a uv / bun / Modal catalog. Domain LoRA specialist + vs few-shot hosted + ([Domain-jev-maker](https://github.com/help-er/Domain-jev-maker); + independent CLINC gold, not a Jev teacher-copy; + matched-precision KL 0.168 vs 0.580 banking; few-shot + determinate McNemar n.s.; train when downstream reads p). + Decide→policy→LLM leftover cascade + ([jav-email-cascade](https://github.com/skiingfalcon/jav-email-cascade); + jev vs gen-json vs gen-logprob; Noul 0.5 never rounded; + license null; mock gen-json flat-confidence is *their mock*). + ORDER BY ranking family + ([jev-orderby-bench](https://github.com/yodablocks/jev-orderby-bench); + six gates pass; Score ordinal 0.143 weak link; 53-way 0.99 + tie; calibration ≠ sortable; recodelabs batch-40 fails + ranking). Wire-compat GLiFormer backend + ([jeff](https://github.com/logan-markewich/jeff); typesafe-sdk + drop-in; ~$2.6 vs $15.6 L4 HTTP ~6×; A10G direct ~$0.65 ~24×; + AG News 75.5% vs 90.5%; CPU more expensive; license null; not + a Jev replica). MED: loopback gateway + ([sysone](https://github.com/hraness/sysone); hosted + local + OpenJev/NanoJev/Mini-Jev; no weights; credential from env). + No wrapper. No invented metrics. +- Hourly ~21:39 Boise 2026-09-18 / 03:39 UTC 2026-09-19 fold + (`research/notes.md` §61): Archer still Watch. Architecture + notes, not a pip / npm / bun catalog. Active-learning triage + ([jev-triage](https://github.com/ThyFriendlyFox/jev-triage); + accept / expensive teacher / human; log full distributions; + **do not distill Jev as teacher of record**, ~68% ceiling). + Evidence-packet explorer + ([jev-semantic-explorer](https://github.com/jimmyhealer/jev-semantic-explorer) + / jevex; index-once ask-many; 1/8→6/8 SWE-bench Verified + finish n=8 *theirs*; packet HitFile 0.233 diagnostic). + Meaning-grep + ([jev-semgrep](https://github.com/uehaj/jev-semgrep); AND/OR/NOT + line Nouls; JP↔EN; MIT LICENSE / GitHub NOASSERTION; 0.94/0.98 + *theirs*). Closed-vote CU + ([JevOnly](https://github.com/buluoray/JevOnly); no planner LLM; + 11 steps / 43 calls / ~$0.014 / 17 s *theirs*). Harbor Jev vs + local MLX PCD vs AR JSON + ([system-one-benchmark](https://github.com/mallahyari/system-one-benchmark); + toxic-chat n=50; Jev 84.0% / Brier 0.1096 vs PCD 52% / 0.3884; + O(1) ≠ calibrated Noul; license null). Host-owned product + ([waymode](https://github.com/mossburgh/waymode); app retains + handlers/permissions; 24/26 + 34/36 *theirs*; not a + self-driving proof). OMP/pi fail-open gates + ([omp-jev-extensions](https://github.com/luw2007/omp-jev-extensions); + `jev_acceptance_gate` + `jev_route`; contrast pi-jev-approver + fail-closed). Skip empty jev-compactor / laya-jolt. No wrapper. + No invented metrics. +- Hourly ~22:38 Boise 2026-09-18 / 04:38 UTC 2026-09-19 fold + (`research/notes.md` §62): Archer still Watch. Architecture + notes, not an `omp plugin` / pip catalog. Watch archive path + missing on this VM; receipts from live GitHub. Permission vs + probability + ([omp-greenlight](https://github.com/SemetricLabs/omp-greenlight); + 1,013 calls / 10 sessions; default **40.9%** prompts removed / + **0 of 94** unsafe auto-approvals on labelled corpus; operator + owns thresholds; plugin never self-tunes; not a sandbox; host + deny stays above). Judgment ≠ permission + ([skill-broker](https://github.com/adamjralph/skill-broker); + Hermes pre-agent outline; code owns grants; Jev never grants + access; **not a production recipe**). Eval integrity / + instrument-not-score + ([dinostomp](https://github.com/collapseindex/dinostomp); + FINDINGS 189 / 99 against itself; `dinostomp jev` if-statement + hygiene; ECE 0.062 *theirs* on 24 examples; beside jevals, not + a Harbor taskset). MED: fast-jev-opencode, jev-desktop, + jev-agent-integration, sift, JevExplore. Census: Awesomejev + 488/21644; SemIf 1641 (+13); jevlike 905 (+4); tracker likes + 41 (+1); Laya yes; Blackwood ABSENT; X MCP flapping + (`pages_archived` 0). No wrapper. No invented metrics. +- Hourly ~23:40 Boise 2026-09-18 / 05:40 UTC 2026-09-19 fold + (`research/notes.md` §63): Archer still Watch. Architecture + notes, not a uvicorn / bun / marketplace catalog. Watch + archive path missing on this VM; receipts from live GitHub + + HF. Hunches labeled. Constrained optimizer + S1 features + ([slo-router](https://github.com/zeeshan8281/slo-router); + license null; Jev task/exactness/evidence as features, never + the sole hot-path gate; fail-open local features; same + routes/accuracy; p95 **77.93 → 490.38 ms** *theirs*; eight-row + demo is not a benchmark). Privilege ≠ verdict + ([construct-auto-classifier](https://github.com/godspede/construct-auto-classifier); + Apache-2.0; effect-based shell gate; fast-allow/deny then Jev + Choice + independent risk Nouls; fail-closed; Jev **0** + dangerous / 975; every chat model leaked 16–104; operator-owned + dials). Attention filter / VOI for human review + ([jev-lens](https://github.com/rashedInt32/jev-lens) + + [jev-lens.nvim](https://github.com/rashedInt32/jev-lens.nvim); + never blocks the agent; never edits; never green unless sure). + MED: [sysone-help/sysone](https://github.com/sysone-help/sysone) + (evaluation-model-first TS SDK; **not** hraness/sysone gateway); + [INSTRUCT_JEV](https://huggingface.co/datasets/ctaxnagomi/INSTRUCT_JEV) + (119 rows; 47/51/21; jevals seed); + [swift-jev](https://github.com/ckaik/swift-jev) (LICENSE-only + this pass; not a CLI product). Census: Awesomejev 488/21644; + SemIf **1652** (+11); tracker likes **42** (+1); lastModified + unchanged; Laya yes; Blackwood ABSENT; X MCP flapping. No + wrapper. No invented metrics. +- Hourly ~00:39 Boise 2026-09-19 / 06:39 UTC fold + (`research/notes.md` §64): Archer still Watch. Architecture + notes, not a uvx / pnpm / marketplace catalog. Watch + archive path missing on this VM; receipts from live GitHub. + Hunches labeled. Measurement owns endorsement + ([jev-packs](https://github.com/dtduc-git/jev-packs); + CC0; nine packs `verified` *theirs* on pinned + `jev-1.13.0`; accuracy/ECE/cost/latency; `unknown` + mandatory; named runner jevassert **not released** / 404; + packs without evidence stay `provisional`). Jev supplies + evidence, code owns authority + ([actiongate-jev](https://github.com/omkarghugarkar007/actiongate-jev); + Apache-2.0; deterministic policy owns ALLOW|REVIEW|BLOCK; + positive score never overrides a hard security fail; + fail-closed financial/destructive/credential if Jev is + down; 500-case is label-baseline, not accuracy). Ranking ≠ + calibration + ([does-jev-confidence-mean-anything](https://github.com/Adilmp/does-jev-confidence-mean-anything) + + [jevcal](https://github.com/Adilmp/jevcal); 8,000 + human-annotated judgments; AUC **~0.91**; stated **~75%** + vs human **~10%**; two-parameter recalibration removes + **~96% ECE** without changing rank; never hard-threshold + raw p as a frequency; vendor "calibrated" often means + rank-correlation). MED: + [gqgs/laya-onnx](https://github.com/gqgs/laya-onnx) + (complete Laya→browser int8; distinct from Mattepiu); + [kunchenguid/local-jev](https://github.com/kunchenguid/local-jev) + (ModernBERT local approximation — not equivalence). Do + not re-fold sysone-help/sysone. Census: Awesomejev + 488/21644; SemIf **1660** (+8); jevlike **910** (+5); + TypeAR 9; tracker likes 42; lastModified unchanged; Laya + yes; Blackwood ABSENT; X MCP flapping. No wrapper. No + invented metrics. +- Same-hour remainder ~00:39 Boise 2026-09-19 (`research/notes.md` + §65): Archer still Watch. Do not re-fold actiongate / jev-packs + / sysone-help. Hot-click CU + ([ego-jev](https://github.com/jiangkoumo/ego-jev); MIT; indexed + viewport table → operation+target; code owns observe/execute/ + `--until`; text model only for type; HN 4.9 s vs 9.7 s / wiki + 5.4 s vs 10.1 s *theirs* n=3, high variance, not a bench). + Jev judges relevance, code decides structure + ([jev-compactor](https://github.com/edwardyen724-g/jev-compactor); + MIT; was empty skip §61; never rewrite; regex floor; compaction + fail-open if Jev down, safety fail-closed; 64.5% / 366 ms / + $0.0004 / 0 invented paths / 4 of 4 facts vs Sonnet summary + 96.2% / 1 invented path, one session). Local rules first, + never auto-train on the model's own hides + ([x-reply-filter](https://github.com/zhuyansen/x-reply-filter); + MIT; `rules.js` then batched Nouls; confirm-queue). OpenCode + port already §62: fast-jev-opencode. Census as §64. No wrapper. + No invented metrics. +- Hourly ~01:47 Boise 2026-09-19 (`research/notes.md` §66): Archer + still Watch. Three clusters: **control-plane combinators** + ([decision-combinators](https://github.com/voidning/decision-combinators); + Then/Gate/Vote/Cascade/Weighted; not literal AND/OR; not chat + turns) + **skill VOI** + ([skillranker](https://github.com/Dicklesworthstone/skillranker); + 52★; two-pass + none-of-these; hook **fail-open** — corrects + §7 fail-closed); **eval integrity without leaderboard theater** + ([jev-capability-atlas](https://github.com/Zaious/jev-capability-atlas) + 10★ receipts, type-safe ≠ correct, axis already §49; + [jev-frontier-100](https://github.com/softpudding/jev-frontier-100) + Jev 77.0% vs Qwen3.5 4B/2048 96.7% / 4B off 56.0%, exploratory; + [jev-ood-calibration](https://github.com/scienthoon/jev-ood-calibration) + 900 tickets ECE 0.107 = 4.4× floor, Choice/Score T~3.3 vs + boolean T 0.66, unknowable priority mean p 0.74); **gate + doctrine clone** + ([turnstile](https://github.com/zyphr-labs/turnstile); Apache-2.0; + policy first, Jev remainder, replay; missing Jev → Review); + **MLX one-pass replica economics** + ([jevmlx](https://github.com/bnsd55/jevmlx); 28★; softmax ≠ + Noul; no local leaderboard yet). MED: invalidate (0 of 157 + false invalidations), jev-intent-review (under construction), + prune-review (22-run cost 1.18% with 305% outlier). Census not + re-derived. No wrapper. No invented metrics. +- Hourly ~02:38 Boise 2026-09-19 (`research/notes.md` §67): Archer + still Watch. Do not re-fold the 01:47 list except sibling + contrast. **TLA+ compose with judgment** + ([jev-labs](https://github.com/copyleftdev/jev-labs); MIT; + never confidently wrong; 1,080 golden 0 wrong *theirs* under + chaos, escalate 5%→18% severe; TLC 1,049,750 states / 0 + errors; synthetic, not clinical). **Advance/coverage ledger** + ([seal](https://github.com/Reasonofmoon/seal); MIT; no seal, + no advance; coverage.path auto|code|human|escalate; mint ≠ + product brain). **skill-broker sibling** (outline already + §62; grants in code vs turnstile runtime vs skillranker + advisory). **Sureness** + ([how-sure-is-jev](https://github.com/adarc8/how-sure-is-jev); + MIT; Choice confidence = max_prob; 75/25 → 0.5 vs entropy + 0.19). **JevBench v1.1** + ([jevbench](https://github.com/fstandhartinger/jevbench); + MIT; unofficial; Main Score 0.6/0.2/0.2; Jev 1.13.0 **87.6**; + calibration reported not scored). **CI typed gate** + ([ci-gatekeeper-bot-jev](https://github.com/NemanjaManic/ci-gatekeeper-bot-jev); + package.json MIT / GitHub SPDX null; 504–629 ms *theirs*). + **Codex MCP adapter** + ([jev-in-codex](https://github.com/teempai/jev-in-codex); + MIT; ranking unbenchmarked; lexical fallback). Census: + SemIf **1683** (+11); jevlike **923** (+5); tracker likes + **43** (+1); Awesomejev 488/21644 unchanged. No wrapper. No + invented metrics. +- Hourly ~03:38 Boise 2026-09-19 (`research/notes.md` §68): Archer + still Watch. Do not re-fold the 02:38 list except sibling + contrast. **Judgment as attention redirect, not a merge + blocker** + ([jev-preflight](https://github.com/muse0509/jev-preflight); + Go MIT; eight risk axes; assist=one reinspect; fail-open; + uncalibrated 0.85; owner-run Claude Code 2.1.267: no-key + fail-open PASS, key-enabled exactly one continuation). + **Landed-script trust / headless≠auto-approve** + ([construct-auto-classifier](https://github.com/godspede/construct-auto-classifier) + delta; cert still Jev **0** dangerous / 975; $0.047/1k). + **Jev judges relevance; code decides structure** + ([jev-compactor](https://github.com/edwardyen724-g/jev-compactor) + product-arm **73%** / 350 ms / 4 of 4 *theirs*; 30–250× + cheaper than shipped summarizers; §65 64.5% is vs-Sonnet). + **Compress-before-first-send** + ([dizk/jev-lens](https://github.com/dizk/jev-lens); MIT; + 79% fewer tokens / 500 SWE-rebench; post-send prune +17% + cost; distinct from rashedInt32/jev-lens). **tools≠use** + ([jev-carryforward](https://github.com/Dharundp6/jev-carryforward) + 0/4 recall; SessionStart > hoping). **Observational + memory** + ([pi-observational-memory-jev](https://github.com/willfish/pi-observational-memory-jev); + keep/kind verbatim; model-free compact). **Independent + open-Jev class** + ([openvons](https://github.com/genai-craft/openvons); + Apache-2.0 LICENSE / GitHub SPDX NOASSERTION; 7★; JevPick + 3.2–4.8×; `/v1/systemone` wire-compat ≠ replica). + **Physical-world S1** + ([HA-Jev](https://github.com/AboveColin/HA-Jev); MIT; + **17★**; sensors from typed answers; not for + locks/heaters). **Judgment outside the store** + ([jevql](https://github.com/kylemclaren/jevql); CLI + judges; vanilla Postgres never sees `jev()`). Short + consumer bullet: [sift](https://github.com/bohutang/sift) + ~$0.00003/post. Census: Awesomejev **561** (+73, agent + tooling 87→107); SemIf **1704**. No wrapper. No invented + metrics. +- Hourly ~04:39 Boise 2026-09-19 (`research/notes.md` §69): Archer + still Watch. Do not re-fold the 03:38 list except sibling + contrast / combinators rename. **Digital-design combinators** + ([jev-combinators](https://github.com/voidning/jev-combinators) + is the rename of decision-combinators; extended Router / + Loop / Retry / Fallback / Memory; metaphor ≠ literal AND/OR). + **VOI cache admission** + ([jevcache](https://github.com/kushals256/jevcache); MIT; + same-intent skip LLM; n=100 *theirs* 0 FP / precision 1 / + recall 0.38 / fpr 0 vs Jaccard@0.35 fpr 0.48; fail-open). + **Harbor skill-routing harness** + ([pi-jev-skill-bench](https://github.com/iamdin/pi-jev-skill-bench) + + [pi-jev-skill-suggestion](https://github.com/iamdin/pi-jev-skill-suggestion); + BM25 vs Jev at roster 50–500; 43 gold; no live numbers this + pass; no-key no-op; tool mode is tools≠use cousin). + **Zeroshot vs BERT displacement** + ([jev-zeroshot-vs-bert](https://github.com/zhuyansen/jev-zeroshot-vs-bert); + +0.05–+0.13 vs DeBERTa-c; contamination 0.901 vs `-c` 0.763; + ≈230 / >2048 labels; DiD 0.035 vs 0.112 *theirs*). + **Typed escalate/continue/abort baton** + ([jev-handoff](https://github.com/shitianfang/jev-handoff); + MIT; inverted loop; gate never grants; fail-open; Vercel + drops confidence). **Worth-your-attention VOI** + ([ThinkyMiner/Winnow](https://github.com/ThinkyMiner/Winnow); + MIT; 80%/90% *theirs*; **≠** kevinpita/winnow). + **Jev WHETHER / Python HOW / LLM WHAT** + ([hermes-jev-router](https://github.com/rsdkrasen/hermes-jev-router); + license null; community plugin; skip-next needs core patch). + **Conflict ≠ ignorance** + ([jev-typed-evaluation-collapse](https://github.com/mleyvaz/jev-typed-evaluation-collapse); + Noul collapses; named Choice p=1.0; binary red 0.67–0.85 + *theirs*). **Playwright executes, Jev chooses** + ([browser-jev](https://github.com/DowLucas/browser-jev); + license null; sample-from-distribution). **Local class** + ([OpenJev](https://github.com/IamBusy/OpenJev) Apache-2.0 + `/v1/decide` 45/60 *theirs*, not TypeSafe drop-in, ≠ + hraness/sysone runners; + [semif-serve](https://github.com/dddanielliu/semif-serve) + 1164 vs 178 ms; runoff ≠ softmax; wire-compat ≠ replica). + Toolbelt notes: jev-security-scan / jev-decisions / TeoMastro + (summary.md 404 this pass); **rh-guard owns reward-hack**. + Flywheel: + [DGUI_HYPERMEM-JEV](https://huggingface.co/datasets/ctaxnagomi/DGUI_HYPERMEM-JEV) + 6-row schema. Census: Awesomejev **flat 561/27007**; tracker + likes **43→45**; SemIf **1714** (+10); jevlike **926** (+3). + No wrapper. No invented metrics. +- Hourly ~05:46 Boise 2026-09-19 (`research/notes.md` §70): Archer + still Watch. Do not re-fold §50–§69 HIGH except sibling + contrast / jevassert landing / prune-review, intent-review, + laya-jolt, local-jev deltas. **Record/replay CI LANDED** + ([jevassert](https://github.com/dtduc-git/jevassert); + Apache-2.0; accuracy/ECE/Brier/cost/latency offline from + recordings; exit 0/1/2; McNemar; Action `@v0`). + **Evidence-gated packs now have a runner** + ([jev-packs](https://github.com/dtduc-git/jev-packs); + size 0→458; 2,990-case matrix *theirs*: Jev/Sonnet 5 + accuracy tie Δ≤0.018, Jev better calibrated 7/9, ~250× + cheaper; sms-spam this-pass 0.953/ECE 0.040). + **Failure-finding arena** + ([jevarena](https://github.com/chenmingtang830/jevarena); + Apache-2.0; **≠** meetr1912/jev-arena; harness not findings). + **BBQ stereotype/uncertainty/cost** + ([jev-bbq-experiment](https://github.com/simonmesmith/jev-bbq-experiment); + license null; 58,492; 97.28%; bias 0.04/0.34; $0.3429 / + 7.75 min *theirs*; not a general bias cert). + **Decider ≠ executor** + ([jeffrey](https://github.com/thomasbrueggemann/jeffrey); + MIT; Jev next-tool/progress/risk/done; LLM fills args; + pick ≠ fill). **Sentence-as-rule lint** + ([jevlint](https://github.com/mizchi/jevlint); MIT; + ast-grep × `ask:`; 13/15 1.00/1.00 *theirs*; **≠** + huntedman/JevLint). **VOI hunk prune** + ([prune-review](https://github.com/shubhangi013/prune-review); + 22-run 1.18% with 305% outlier; ~20% target; cost not + quality). **Whole-repo intent** + ([jev-intent-review](https://github.com/yottayoshida/jev-intent-review); + VERIFIED/VIOLATION/UNKNOWN; empty search ≠ proof). + **GLiNER2 System One spec** + ([Jev_from_GLiNER2](https://github.com/Eran-BA/Jev_from_GLiNER2); + spec-only; ≠ jeff). **Open replica substrates** + ([grande](https://github.com/bokuweb/grande) JGLUE 0.614/ + 0.853 + 270M 0.710/0.710 *theirs*; + [laya-jolt](https://github.com/jlt-commons/laya-jolt) + byte parity; [JEV-CPU](https://github.com/leesk212/JEV-CPU) + PoC, Meanblock 404; [local-jev](https://github.com/kunchenguid/local-jev) + done 30%/shape 57%). **Persist constraints** + ([pi-heed](https://github.com/Nyarlathoteppppp/pi-heed); + 98.5%/0 false block *theirs*). Toolbelt note: + actiongate slogan already §64. MED: + [system-one-responsible-ai](https://github.com/david-j-lustig/system-one-responsible-ai) + size-0 framing stub. Census not re-derived. No wrapper. + No invented metrics. +- Hourly ~06:43 Boise 2026-09-19 (`research/notes.md` §71): Archer + still Watch. Do not re-fold §50–§70 HIGH except sibling + contrast. **Harbor SGR-judge contract** + ([jev-judge-bench](https://github.com/slavadubrov/jev-judge-bench); + README MIT / GitHub SPDX NOASSERTION; frozen SLA-150; Jev vs + Luna / DeepSeek-flash / glm-5.3-flash; invalid = FN; + 21 offline tests; canaries ≠ quality; **no quality headline + yet**; **≠** jevarena / jevbench). **Empty skip** + ([jev-context-pruner](https://github.com/IPECTER/jev-context-pruner); + 409 empty). **Hand no-text steps** + ([jev-use](https://github.com/shitianfang/jev-use); MIT + v0.4.1; p50 220 ms; 186 vs 2,672 ms; gate 12/12; Vercel + drops confidence → margin 0.4; first loop 17/20 then 0/20 + *theirs*; **≠** jev-ultrafast). **Pi System-One control + plane** ([pi-jev-control](https://github.com/goodruizhan/pi-jev-control); + license null; v0.3.0 private; GUI never force-click; + compaction never writes session). **Never free-generates** + ([jev-gpt](https://github.com/florian-hoenicke/jev-gpt); + license null; ~400 calls / 75 s / 2¢ *theirs*). + **OpenRouter recipe atlas** + ([jev-cookbook](https://github.com/nexibeo/jev-cookbook); + MIT; 1★; 16–36 samples not benches; 425 calls / $0.015; + browser 5/6 *theirs*). **Personal-history feed** + ([jevfeed](https://github.com/fengyiqicoder/jevfeed); MIT; + no social graph; one request per batch of ten). + **Competing NAR claim-audit, not endorsement** + ([openJev-verdict-2.0](https://github.com/Heman10x-NGU/openJev-verdict-2.0); + README Apache-2.0 / GitHub SPDX NOASSERTION; 77.10%/0.0636/ + 0.0144 *theirs* unverified; **open PR #1**: throughput≠ + latency, Laya parity, like-for-like ECE; **≠** + IamBusy/OpenJev). Census not re-derived. No wrapper. No + invented metrics. +- Hourly ~07:49 Boise 2026-09-19 (`research/notes.md` §72): Archer + still Watch. Do not re-fold §50–§71 HIGH except sibling + contrast. **1-token logprob endpoint ≠ Noul** + ([chakuho](https://github.com/taku-me/chakuho); MIT; + coverage ≠ correctness; GUI 336 *theirs* 27B 95%/92% + vs Jev 89%/82%; `__none__` 97% vs 8B 10%). **Open + replica engine** ([jevinf](https://github.com/zerodegress/jevinf); + MIT; 2.57×/2.27× 100% argmax; MPS only). **Unofficial + Elixir SDK ≠ OTP peer** + ([typesafe-elixir-sdk](https://github.com/phiat/typesafe-elixir-sdk); + MIT; 1★; ≠ dannote/jev). **jevex rename + n=16 VOI** + ([jevex](https://github.com/jimmyhealer/jevex); 160s→69s + / $8.74→$3.13 / 16/16 *theirs*; keep n=8 1/8→6/8). + **Commit attention≠verdict** + ([commitjev](https://github.com/yodablocks/commitjev); + MIT; middle band never rounded; 0 false on 5 clean + *theirs*). **Hermes plugin is Agnes not TypeSafe** + ([hermes-plugin-jev](https://github.com/Mrmimee/hermes-plugin-jev)). + **Pi compact ≠ compaction** + ([pi-jev-compact](https://github.com/dev-willbird1936/pi-jev-compact); + MIT). **Empty skip** + ([jev-runway](https://github.com/IPECTER/jev-runway); + LICENSE-only). **Decision-native inbox** + ([mailordinal](https://github.com/Milo318/mailordinal); + MIT). **Unofficial jev-cli not ready** + ([jev-cli](https://github.com/shaharia-lab/jev-cli); + 0.0.0; ≠ jevql). **Laya multilingual** + ([laya-multilingual](https://huggingface.co/convaiinnovations/laya-multilingual); + MASSIVE 0.366/0.387; Khmer 0.000@0.952; ships + uncalibrated). **Schema-scorer Hub** (GitHub 404; v2 + Choice 0.841; peaked ranking). HF 401 this pass on + open-jev-laya-bench / jev-tree-choice-cap / + INSTRUCT_JEV; jevlogs 404+401. Census not re-derived. + No wrapper. No invented metrics. +- User-provided signal ~08:37 Boise 2026-09-19 + (`research/notes.md` §73): **Skip Archer.** + **Productized System One HTTP** + ([classifier-dev](https://github.com/mrmps/classifier-dev); + MIT; **185★**; https://classifier.dev). Label + + calibrated confidence as the public contract; batch + `{id,text}[]` ~1000; Jev primary, LLM fallback only. + 400 headlines **650 ms** *theirs*. **Escalate-under- + threshold:** smart re-asks single-label <0.7; + multi-label ignores (re-judge worse, 23 s). Emotion + ≥0.9 → 82% / <0.5 → 29%; gemini-3.8-flash 87.5→90.0 / + 61.8→63.7 *theirs*. Multi-label F1 **0.887** / **230 ms** + vs cascade **0.799** / 1.5 s (eval 232 ms; AG News + **87.7%** vs 82.0%). **Measurement-first:** `/benchmark` + from tracked JSON; read eval/README (n=7 train-on-test; + ~0.03 coin flip). **Silent FALLBACK:** granite F1 + **0.546** vs advertised ~**0.800** *theirs*; rh-guard + owns the gate. Life/business (spam/inbox/feedback), + not SWE-only. Distinct from ask-jev-ai wall. No wrapper. + No invented metrics. +- User-provided signal ~08:48 Boise 2026-09-19 + (`research/notes.md` §74): **Skip Archer.** **Delta of + §48.** Pointer-not-generator at evidence-synthesis + scale + ([choxos/jev-reviewer](https://github.com/choxos/jev-reviewer); + MIT; **12★**; https://jevreviewer.xera.ac). **≠** + [egma-ai/jev-reviewer](https://github.com/egma-ai/jev-reviewer). + Two-pass Choice (which line) + Noul (does this line + itself answer); quotes = Noul ≥ 0.5 *theirs*. *Not + found* / *Unclear* first-class. Human tick is the + product (checked never overwritten). 18-q template + **4.6 s / $0.0101** *theirs* (spot check, not a + validation study). Cochrane / PRISMA / RoB, not + SWE-only. No wrapper. No invented metrics. +- User-provided signal ~08:56 Boise 2026-09-19 + (`research/notes.md` §75): **Skip Archer.** + **Wire-compat ≠ logit-equiv** + ([githubnext/localjev](https://github.com/githubnext/localjev); + MIT; **261★**; GitHub Next). **≠** + [kunchenguid/local-jev](https://github.com/kunchenguid/local-jev). + Bun `POST /v1/systemone` on DiffusionGemma via Chat + Completions; TypeSafe SDK drop-in. Prompted JSON → + validate/retry → normalize + entropy confidence — not + razorback16 structured-read logits. Harbor-shaped + bake-off *theirs*: 1,200 req; Qwen3.6 short macro + **76.7%**; Gemma 4 26B-A4B **75.0%**; DiffusionGemma + **74.2%**; no definitive winner (2/120); do not treat + as calibrated. LM Studio still cannot load + DiffusionGemma. Do not copy bun / `.env`. No wrapper. + No invented metrics. +- User-provided signal ~09:07 Boise 2026-09-19 + (`research/notes.md` §76): **Skip Archer.** **Laya + packaging, not a new species** + ([NandhaKishorM/laya](https://github.com/NandhaKishorM/laya); + Apache-2.0; **710★**). PyPI + `Router` over Hub + [`laya`](https://huggingface.co/convaiinnovations/laya) / + [`laya-multilingual`](https://huggingface.co/convaiinnovations/laya-multilingual) / + [`laya-typed-decisions`](https://huggingface.co/convaiinnovations/laya-typed-decisions). + **≠** TypeSafe `/v1/systemone`. **≠** githubnext/localjev. + T4 *theirs*: 1q **32.8 ms** (~7.8× vs Jev p50 + 236–276 ms). Post-T ECE **0.081** vs Jev **0.246**; + raw ECE still trails (0.213 vs 0.144). Banking77 + **0.425** vs Jev **0.870** (77 vs 72; ~3–4 tok/label). + typed-decisions **0.766** is a fine-tune (base + 0.362/0.342 vs majority 0.461). Soft-acc 0.471 vs + 0.580. Khmer **0.000@0.952** — Router because gating + cannot catch. 0.85 still soft. Jev rows third-party + unpublished-here. Do not copy pip / preload. No + wrapper. No invented metrics. +- User-provided signal ~09:14 Boise 2026-09-19 + (`research/notes.md` §77): **Skip Archer.** **External + openjev census ≠ scored bake-off** + ([@airesearch12](https://x.com/airesearch12/status/2101259522933186879); + Florian S / Benchmark Heaven). Named ~18 (system-one-open, + openjev-sglang, DeBERTa open-jev, Needle 3, + open-alternative-jev, Nimble 9B, SemIf, open-jev Dasein / + JoshuaSP, OpenJev razorback16, mini-jev, system-one, + system-one-gemma, jevlike, AlexWortega/openjev, GLiNER2, + Succinct Router 14M, jev-model-router/Director/Loki). + GLiNER2 + routers are **class-boundary**. Incomplete vs + Laya / localjev / kev / TypeAR / openvons. Engagement + **ephemeral** (SIGNAL ~417/9/3; this pass 564/15/5). **≠** + jevbench v1.1. Watch + [jev-models](https://benchmarkheaven.com/jev-models); do + not paste live ranks. Do not copy Stripe. No wrapper. No + invented metrics. +- User-provided signal ~09:24 Boise 2026-09-19 + (`research/notes.md` §78): **Skip Archer.** **JevBench v1.2 + scored board** + ([benchmarkheaven.com/jev-models](https://benchmarkheaven.com/jev-models); + harness [fstandhartinger/jevbench](https://github.com/fstandhartinger/jevbench) + MIT; 0★; HEAD `27ed3d6c`). Protocol `jevbench::v1.2`; scored + 19 Sept 2026; 534 decisions (hard 220 = 30% of Intelligence). + Official Score = geometric mean of I/C/S/K at 25% each. Jev + 1.13.0 **75.3**; SemIf (Qwen3.5-4B) **74.6** (−0.7); OpenJev + DiffusionGemma (razorback16) 67.6 *theirs*. Luna Intelligence + **96.8** rank **#7** on cost. Calibration **on** the rank + (delta from v1.1). Weighting is a product design. Option-order + 72%→21%. Self-host latency ×2 is an assumption; many costs + est. Laya absent (gap, not named-excluded). GLiNER2 mapping + issues; apps out. Qwen3.8 27B Chutes TEE **≠** Archer. **≠** + tweet census §77 **≠** v1.1 87.6. Do not copy Stripe / CLI. + No wrapper. No invented metrics. +- Hourly System One watch ~08:42 Boise 2026-09-19 + (`research/notes.md` §79): **Skip Archer.** Named HIGHs + **already folded** (§73–§78) — extract **how-to-apply**, + not a hit list: wire-compat ≠ logit-equiv (prompted JSON + ≠ structured logit); productize label+p and mark + `FALLBACK`; packaging ≠ new species / script-before-p / + 0.85 still soft; pointer-not-generator (two-pass; *Not + found*; human tick); external census ≠ scored bake-off / + geo-mean weights are a design. **Skip thin noise** + (JEValuate / jevspeak / fable-jev; jev-semgrep already + §61). Hard-gating a Noul as a PR/quality gate is + soundness theater + ([totally-tim/jev-gate](https://github.com/totally-tim/jev-gate) + 0★ ≠ jev-gateway; [claude-jev-warden](https://github.com/connectedGraph/claude-jev-warden) + 1★). Qwen3.8 27B ≠ Archer. No wrapper. No invented + metrics. +- User-provided HIGH ~09:50 Boise 2026-09-19 + (`research/notes.md` §80): **Skip Archer.** Delta of + §46, not a new species. + [khordoo/jev-reflex-autonomy-lab](https://github.com/khordoo/jev-reflex-autonomy-lab) + (TypeScript; **7★**; license null; README SHA + `130987c9`; ARCHITECTURE SHA `48da0769`; HEAD + `e3297ebe`). S1 never stalls waiting; S2 is one-use + advisory and never flies. Purple confidence = + **consumed** S2 (purple bar = arrival; red = fail). + Local controller is rule-based **≠** + githubnext/localjev **≠** kunchenguid/local-jev. + Live API `POST /v1/systemone` `jev-latest`; 20% + starting gate *theirs* still soft and does not start + a mission. Seed = geometry ≠ async replay. No pixels + to either provider; confidence ≠ selected + probability; S2 never grants. README GLM 5.3 vs + ARCHITECTURE muse-spark-1.3-contributor — quote both + *theirs*. Experimental viz, not a flight controller. + Do not copy npm / `.dev.vars`. No wrapper. No + invented metrics. +- User-provided HIGH ~09:51 Boise 2026-09-19 + (`research/notes.md` §81): **Skip Archer.** Productized + observe→score-among-candidates→code-acts on a Mac, + not a new species. + [awlevin/typesafe-computer-use](https://github.com/awlevin/typesafe-computer-use) + (Python; MIT; **427★**; README SHA `369f4a6a`; HEAD + `cc7b5066`). OCR+AX → numbered items → TypeSafe + Choices (`kind`/`item`/`site`/`offscreen`) → + deterministic click/type. **Never ships a screenshot + to frontier for the *decision***; the one-shot + **answer** writer may receive the capture (reader + packet, not the Choice). Writer only for free text. + Overlapping options = false low confidence. AX bonus + never sole (Spotify 0 *theirs*). Post-type Noul 0.5 + and `--min-confidence` 0.4 still soft. $0.0002 vs + Opus $0.032 (155×) *theirs* on **one screenshot**, + not a Harbor taskset. Honest caveat: dates.py rebuilds + pixel-free reasoning. **≠** jev-ultrafast **≠** + cua-s1 **≠** jev-macos-loop **≠** camoufox. Do not + copy `uv sync` / `.env`. No wrapper. No invented + metrics. +- User-provided HIGH ~10:01 Boise 2026-09-19 + (`research/notes.md` §82): **Skip Archer.** Productized + ASR observe→score-among-candidates→code-acts in + headed Chromium, not a new species and not omni. + [moritzkremb/jev-voice-browser](https://github.com/moritzkremb/jev-voice-browser) + (JavaScript; MIT; **103★**; README SHA `fa033303`; + HEAD `054db0f3`). Web Speech partials → one 9–11- + question Jev request (~250–350 ms *theirs*) → + policy. Pointer-not-generator for spans. Closed-set + may act on a partial; free-text waits. Spoken + confirm is convenience, not auth. Numbered overlay, + no second model. Integration 27/27 / ~$0.0002/call + *theirs* fixtures, not a Harbor taskset. 0.5 / 0.55 + / 0.6 still soft. **≠** jev-voice-control **≠** + nikolas-j **≠** Aj1905 **≠** typesafe-computer-use + OCR. Do not copy `npm` / `.env` / `run.sh`. No + wrapper. No invented metrics. Expand the §39 tweet; + do not re-card it. +- User-provided HIGH ~10:20 Boise 2026-09-19 + (`research/notes.md` §83–§84): **Skip Archer.** Two + signals, one fold. + [reddpy/AgentGhost](https://github.com/reddpy/AgentGhost) + (TypeScript; MIT; **2★**; README SHA `44145fa9`; + HEAD `ac04e4fb`). Intent-aware ALLOW/ASK/DENY + wrap-as-execution: the wrap *is* the tool function; + rules first; ASK throws; `failMode: closed`. Judge + is a slot. Provider-hosted tools out of reach. + **≠** jwen5419807/agentghost **≠** vventirozos + **≠** actiongate **≠** toolgate **≠** jev-use. + rh-guard owns the gate cousin. Do not copy `npm` / + `.env` / `AUTO_APPROVE`. [@studio_yebisu JP genre + atlas](https://x.com/studio_yebisu/status/2101065176069886152) + (2026-09-18T21:45:48Z). Apps by hole, not a scored + bake-off. Stars research-time (typesafe-computer-use + 203→**427**; jev-voice-browser 40→**103**). Not + verified evals. Engagement ephemeral (this pass + 131,234 / 1,934 / 192). SAM 3.1 already §39. + OpenRouter Jev no-waitlist is WATCH, not a recipe. + **≠** @airesearch12 class census **≠** v1.2 board. + Do not dump the 30 repos. No wrapper. No invented + metrics. +- User-provided HIGH ~10:25 Boise 2026-09-19 + (`research/notes.md` §85): **Skip Archer.** External + pedagogy / how-to-apply, not a new species. + [@akshay_pachaar “Jev Clearly Explained”](https://x.com/akshay_pachaar/status/2101037514945597645) + (article https://x.com/i/article/2100940576741093376; + 2026-09-18T19:55:53Z). LLM hammer for bounded + decisions; code owns branches; parallel questions; + thresholds in code; **schema-safe ≠ correct**; + placements = routing / tool-risk / verify with LLM; + shadow-mode; questions-as-code. **200× / 400×** and + 70–500 ms / $0.042/MTok are TypeSafe **ceiling** + claims *theirs*, not Harbor. Text-only; not looking + at the screen. **≠** official docs **≠** Flavio + Copes **≠** LangChain harness **≠** AgentGhost. + Engagement ephemeral (this pass 233,495 / 2,280 / + 235). Do not copy the Python samples. No wrapper. + No invented metrics. +- User-provided HIGH ~10:30 Boise 2026-09-19 + (`research/notes.md` §86): **Skip Archer.** Dedicated + fold of [uehaj/jev-semgrep](https://github.com/uehaj/jev-semgrep) + (light-noted §61). Grep by meaning via Jev Noul; + proposition ≠ embedding; contrast-set (all six + about a refund; only customer-*asking* pass); + AND/OR/NOT are boolean ops on *thresholded* bits + (do not multiply p; ≠ jev-combinators metaphor). + Cross-lingual; no index; EN safer near threshold. + Semgrep.dev SAST name collision. **Not a gate** + (ranking fail-open; rh-guard skip). LICENSE MIT / + GitHub NOASSERTION. HEAD `21120e9`; README SHA + `923e6a5`. Stars ephemeral (0 → SIGNAL ★42 → **51** + this pass). 0.94/0.98 LLM-as-judge 10×51 *theirs*, + not Harbor. **≠** jevgrep **≠** jev-sift **≠** + jevex **≠** semgrep.dev. Do not copy npm / `npx` / + `.env` / marketplace. No wrapper. No invented + metrics. +- Hourly 1047 HIGH + deferred 0945 backlog + (`research/notes.md` §87): **Skip Archer.** Docs-only + off main (PR #2 merged). How-to-apply / mental + models / architecture / Harbor-jevals — not + SWE-only. Formal methods compose with scoring; a + Noul is a SENSOR; hard-gating as test/PR/HA + write/authorship is soundness theater. Ten + clusters: decision-validated UI + ([gram-render](https://github.com/wei-b0/gram-render) + never authors text; + [jev2ui](https://github.com/dglazkov/jev2ui) Jev + decides / Gemini writes); + decision-as-assert + ([jevtest](https://github.com/realZachi/jevtest) + ambiguous band 0.15–0.85; 0.85 still soft); + hybrid S1 + ([anima3](https://github.com/hulryung-uo/anima3) + Qwen logprob default; jeff confidently flat; do + **not** invent Laya); + pointer search + ([JevFind](https://github.com/Peu77/JevFind)); + Harbor trio + ([jev-frontier-bench](https://github.com/manjunathshiva/jev-frontier-bench) + 72.5%/ECE 0.161 vs Fable 84%/0.064 *theirs*; + ChaosNLI JS worse than uniform; **≠** + frontier-100; + [jev-gliclass-bench](https://github.com/JoeSlain/jev-gliclass-bench) + product bakeoff 78/40/49; + [job-posting-triage](https://github.com/geckguy/job-posting-triage) + majority floor 0.947 / tfidf wins / calibration ≠ + discrimination); + authorship named escape (not evidence); + non-SWE + ([ha-switchboard](https://github.com/grayslawson/ha-switchboard) + HA remains execution **≠** HA-Jev; + [n8n-nodes-jev](https://github.com/vibe-with-me-tools/n8n-nodes-jev) + unofficial Low Confidence); + compaction delta + ([fast-jev-compaction-pi](https://github.com/zaycruz/fast-jev-compaction-pi) + **≠** pi-jev-compact **≠** pi-jev-compaction; + ~50× *theirs*); + full-distribution optimizer + ([jevloop](https://huggingface.co/spaces/async-dime/jevloop) + UCB1+CEM; no LLM in the loop; mock default); + deferred class + ([laya-vision](https://huggingface.co/thaitea/laya-vision-smolvlm-256m) + SmolVLM; `score` untrained; **≠** blackwood **≠** + Archer; + [Cerebellum-2B](https://github.com/mkeco/Cerebellum-2B) + `/v1/decide` ≠ TypeSafe; wire-compat vs + agent-routing as separate Harbor axes; competing + NAR **not endorsement**; + [laya-grounded](https://huggingface.co/Luni/laya-grounded) + not drop-in; phishing 0.611→0.512; Platt not + temperature). 0★ HIGHs still get real cards. Do + not copy npm / `pi install` / n8n / HA add-on / + `TYPESAFE_API_KEY`. No wrapper. No invented + metrics. +- Queued user SIGNALs + remaining deferred 0945 HIGH + (`research/notes.md` §88): **Skip Archer.** Docs-only + on PR #3. Mental models: open LoRA replica + ([GestaltLabs/Jeff-1](https://huggingface.co/GestaltLabs/Jeff-1) + acc **0.8183** ECE **0.0807** vs Jev **0.8283** / + **0.0932** n=9730 *theirs*; set reused; **≠** + [logan-markewich/jeff](https://github.com/logan-markewich/jeff)); + Jev-first bounded agent + ([stanley-code](https://github.com/devagrawal09/stanley-code) + empty findings ≠ approval; human promote; 0.6/0.55/0.15 + still soft); + NL memory → beam-search FS + ([findme](https://github.com/marc2332/findme) **≠** + JevFind); + price workers not the conversation + ([jevsubrouter](https://github.com/leftspace89/jevsubrouter) + fail-open; counts ≠ dollars). laya-vision + + Cerebellum already §87 — not re-carded. Soft Noul ≠ + hard safety. Do not copy `uv` / npm / cargo / + marketplace / `TYPESAFE_API_KEY` / `JEVSUB_API_KEY`. + No wrapper. No invented metrics. +- Hourly 1144 HIGH (`research/notes.md` §89): **Skip + Archer.** Docs-only on PR #3. Do **not** re-fold + 1047 / §87 / §88. How-to-apply / mental models / + architecture / Harbor-jevals — not a thin Jev skill + dump. Backend-agnostic categorization/scoring/ + decision class. Formal methods compose with scoring; + a Noul is a SENSOR; hard-gating a default 0.5 bool, + quoting apa “mathematically fulfilled,” treating A/B + proxies as token savings, letting Jev send mail, or + pasting jev-test bars as results is soundness theater. + Seven clusters: **Typed if** + ([feelings](https://github.com/BoundaryML/feelings) + `.feels()` default 0.5 is Noul-0.5-never-rounded; + exhaustive BAML `match`; **≠** hunch **≠** Probably; + license null; **0★**); + **Shadow then honor** + ([apa-agent-harness](https://github.com/AiPersonacademy/apa-agent-harness) + **≠** AntonioCoppe/jev-harness; unpublished npm; + 0.85 still soft; + [apa-persona-engine](https://github.com/AiPersonacademy/apa-persona-engine) + SM then leftover LLM; <250 ms ≠ microsecond; + [grok-bot-jev](https://github.com/Bodila51/grok-bot-jev) + skill honor; A/B proxies ≠ tokens; 13.0× is a + top-five cap); + **Human every action** + ([Essentiel-Jev](https://github.com/JacquesGariepy/Essentiel-Jev) + never authority; 0.75 provisional; license null); + **Atom then sense** + ([enzo-mcp](https://github.com/mahawi1992/enzo-mcp) + independently falsifiable claims; UNKNOWN useful; + **≠** jev-sift); + **File by Choice** + ([pigeonhole](https://github.com/noripto/pigeonhole) + OTHER skip; 0.6 still soft; **≠** jev-semgrep; + client-side playground + [jev-agent-decision-playground](https://huggingface.co/spaces/bojansandhaus/jev-agent-decision-playground) + static no-network; **≠** classifier.dev; sibling + jev-decisions pointer only); + **Question preflight** + ([jev-reliability](https://github.com/vcjdeboer/jev-reliability) + Nothing about accuracy; noul-gate 0.0%/12.5%/3.6% + *theirs*; **≠** dinostomp; + [clduab11/jev-test](https://github.com/clduab11/jev-test) + bars ≠ scores; **≠** realZachi/jevtest; + [jev-rag-benchmark](https://github.com/erendikmenn/jev-rag-benchmark) + “Jev wins” is not an assumption; **≠** Jev-RAG; + [dairui1/jev-lab](https://github.com/dairui1/jev-lab) + urgent 91% vs Haiku 79% *theirs* synthetic; **≠** + BrendanH18/jev-lab; do not re-card jev-desktop); + **Inbox read-only vs write** + ([jevmail](https://github.com/fazlerocks/jevmail) + `gmail.readonly` ~3¢/1k *theirs*; **3★**; + [mailjay](https://github.com/secondfret/mailjay) + archive/trash after review; license null; **≠** + mailordinal). Soft Noul ≠ hard safety. 0★ HIGHs + still get real cards. Do not copy unpublished npm + `@aipersona/…` / `uv` / `baml toolchain` / Gateway + keys / `TYPESAFE_API_KEY`. No wrapper. No invented + metrics. +- Hourly 1241 HIGH (`research/notes.md` §90): **Skip + Archer.** Docs-only on a **fresh PR off main**. Never + reopen merged #3 / #4 / #5. Do **not** re-fold 1144 / + §89. How-to-apply / mental models / architecture / + Harbor-jevals / toolbelt — not a thin Jev skill dump. + Backend-agnostic categorization/scoring/decision class + (Jev-like speed/econ). Formal methods compose with + scoring; a Noul is a SENSOR; hard-gating AUTO_ACT, + treating ranking logits as frequencies, pasting + 95.2% / 83% plumbing / 36/120 NL2Bash as class + ceilings, or letting Jev send/delete/close pinned + tabs is soundness theater. Twelve clusters: + **Observe→score→act namesake** + ([ZHUBoer/ego-jev](https://github.com/ZHUBoer/ego-jev) + reserved `__none__`; runWorkflow completed ≠ success; + **≠** jiangkoumo/ego-jev; **0★**); + **Decision-as-ranking** + ([jsort](https://github.com/keltokhy/jsort) scores are + relative; Noul not Choice for scale; CommonLit + r=0.824 / ρ=0.841 *theirs*; **1★**); + **Native vs schema-guided Harbor** + ([groundedness-judge-bench](https://github.com/slavadubrov/groundedness-judge-bench) + native vs schema-guided; implicit_true included + in yes; Jev 0.6667 vs GLM 0.7661 *theirs*; + LICENSE MIT / SPDX NOASSERTION; **0★**; **≠** + jev-judge-bench); + **0 promotions / authored vs real** + ([jev_playground](https://github.com/JYeswak/jev_playground) + 0 promotions; routing-backtest 0.0447%; **0★**; **≠** + HF playground); + **Offload + classifier-not-generator** + ([yuyang2230/jev-agent-skill](https://github.com/yuyang2230/jev-agent-skill) + jev-1.13-free; **≠** GodsBoy; **0★**; + [jev-techstack-classifier](https://github.com/swap-mitra/jev-techstack-classifier) + stack_config.json only; license null; **0★**); + **Collapse late** + ([s1_ruby](https://github.com/innocentdiaz/s1_ruby) + collapse late; `undecided?` abstain; **≠** hunch **≠** + feelings; **1★**); + **Unofficial toolbelt** + ([2389-research/judgement](https://github.com/2389-research/judgement) + license null; confidence ≠ winner p; **0★**; + [typesafeai-sdk-rust-community](https://github.com/community-ports/typesafeai-sdk-rust-community) + typesafeai-sdk-community not a new species; **0★**); + **Pointer shell** + ([tpellet/hunch](https://github.com/tpellet/hunch) + exit 3; never-execute list; **≠** carldaws/hunch; + **0★**); + **Preview-first VOI / rubric rewrite** + ([jev-file-search](https://github.com/emilwagman/jev-file-search) + scores not calibrated accuracy; **0★**; + [jev-linkmap](https://github.com/stas4000/jev-linkmap) + Jev never sees S2 prose; LICENSE MIT / SPDX + NOASSERTION; **0★**); + **Life fail-open covers** + ([jev-mail](https://github.com/muhammedilyasy/jev-mail) + metadata only; [tidy](https://github.com/MANISH007700/tidy) + none-of-folders stay; [tab-bouncer](https://github.com/MANISH007700/tab-bouncer) + pinned/audio/current never closed; + [lkclean](https://github.com/stefw/lkclean) Show + fail-open; [jev-yt-time-saver](https://github.com/jaibhasin/jev-yt-time-saver) + Show anyway); + **S1 decide / S2 plan** + ([ORIGIN-CIVILIZATION](https://github.com/JacquesGariepy/ORIGIN-CIVILIZATION) + pause-if-no-Jev; validResponse sums-to-1; **≠** + Essentiel-Jev; **1★**); + **Seed/expand/judge/verify + local daemon ≠ Jev** + ([jev-crawlers](https://github.com/russfranky/jev-crawlers) + risk bands never raw boolean; **0★**; + [jevbrain](https://github.com/Synxneuos/jevbrain) + AUTO_ACT is not a Noul; license null; **9★**). Soft + Noul ≠ hard safety. 0★ HIGHs still get real cards. + Census (user-provided; not re-derived): Archer still + NOT landed (HF empty; tracker likes 49 lastModified + 2026-09-19T18:37:18Z still promised); Laya yes; + Blackwood ABSENT; SemIf 1846 (+17); jevlike 962 (+3); + TypeAR-AI/TypeAR 10 (+1); Awesomejev flat 561/27007. + Do not copy `TYPESAFE_API_KEY` / `ZEN_API_KEY` / + `AI_GATEWAY` / `uv` / `cargo` / wrangler / chrome + unpacked / OAuth client ids / shop URLs / `.env`. + No wrapper. No invented metrics. +- Hourly 1347 HIGH (`research/notes.md` §91): **Skip + Archer.** Docs-only on a **fresh PR off main**. Never + reopen merged #3 / #4 / #5 / #7. Do **not** re-fold + 1241 / §90. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class (Jev-like + speed/econ). Formal methods compose with scoring; a + Noul is a SENSOR; hard-gating argmax as safety, + pasting 97.7% n=130 / 0 hallucination / 8,026 tokens + as class ceilings, or treating a skill named System + One as a judge is soundness theater. Nine clusters: + **Judge harness as control API** + ([judgekit](https://github.com/lexingtonhibiki/judgekit) + YAML classify/score/route/verify; 97.7% n=130 + *theirs*; **0★**; + [typed-judge-kit](https://github.com/Ascurse/typed-judge-kit) + verdict-in-code; MIN_LABELS=20; **0★**); + **Batch packing VOI** + ([decide](https://github.com/alsoleg89/decide) + packing VOI; 0.8 ≠ 80% accuracy; license null; + **0★**; **≠** jev-sift); + **Calibration as product** + ([Jev-Calibration](https://github.com/AnthusAI/Jev-Calibration) + Platt ECE 0.117→0.052; license null; **0★**; + [jev-calibration-arena](https://github.com/pmcclelland/jev-calibration-arena) + never acts; size 0; **0★**; **≠** jev-arena); + **Decision-as-Plugin** + ([openrouter-jev-mcp](https://github.com/ctmx/openrouter-jev-mcp); + [typesafe-mcp](https://github.com/cyrusasco/typesafe-mcp) + noul deadband 0.35–0.65; + [FrancoisChastel/jev-code](https://github.com/FrancoisChastel/jev-code) + ≠ npm jev-code; **1★**; + [claudecode-jev-marketplace](https://github.com/skylence-org/claudecode-jev-marketplace) + fail-open not hot path; + [mcp_jev](https://github.com/pedroknigge/mcp_jev) + packs not ask_jev; + [jev-skill](https://github.com/codaaiteam/jev-skill) + jevtypesafeai.com ≠ TypeSafe); + **Policy-constrained skill select** + ([hermes-switchyard](https://github.com/bgrablin/hermes-switchyard) + ≠ hermes-jev-router ≠ hermes-plugin-jev; **0★**); + **Tiny local econ pruner** + ([nanoprune](https://github.com/dmdjr1409/nanoprune) + 2.8MB ECE 2.58%; 0 hallucination theater; **0★**); + **Observe→score→act cousins** + ([jev-browser-agent](https://github.com/smartdio/jev-browser-agent) + ≠ ZHUBoer/ego-jev; + [omp-jev-web](https://github.com/Dakai/omp-jev-web) + DONE ≠ proof; + [hari007sh/jev](https://github.com/hari007sh/jev) + ≠ dannote/jev; license null); + **Deterministic verify ≠ System One** + ([system-one-skills](https://github.com/0thernet/system-one-skills) + deterministic verify; **0★**); + **Soft-score vs hard-argmax** + ([typed-gate](https://github.com/harshpuri84/typed-gate) + band [0.40,0.60] is refusal; + [pi-jev-gate](https://github.com/fivethirty/pi-jev-gate) + fail-closed; choice is the verdict; rh-guard owns). Soft + Noul ≠ hard safety. 0★ HIGHs still get real cards. + Census (user-provided; not re-derived): Archer still + NOT landed. Do not copy `TYPESAFE_API_KEY` / + `OPENROUTER_API_KEY` / `JEV_API_KEY` / `uv` / `npx` + / plugin-marketplace install / `.env`. No wrapper. + No invented metrics. +- Hourly 1441 HIGH (`research/notes.md` §92): **Skip + Archer.** Docs-only on a **fresh PR off main**. Never + reopen merged #3 / #4 / #5 / #7 / **#8**. Do **not** + re-fold 1347 / §91. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class (Jev-like + speed/econ). rh-guard owns the gate cousins; + Augustus owns placement. Formal methods compose with + scoring; a Noul is a SENSOR; pasting Foq 100%/ECE + 0.2% / Reranker 0.1667 / 400 ms / akpsahan vs-Jev as + class ceilings, or treating Qwen3.8-27B as Archer, is + soundness theater. Nine clusters: + **Self-hosted econ** + ([Foq](https://github.com/yohanargentina-oss/Foq) + ~25ms/2.2GB local; **1★**; + [rev](https://github.com/jaswanthsanjay88/rev) + prefill-only + HF jev-0.5b; **0★**; + [robfrase/jev](https://github.com/robfrase/jev) + planning memo); + **Soft-judgment gate integrity** + ([typesafe_agent_gates](https://github.com/ThiagaoBR/typesafe_agent_gates) + 27/27 / 31/31; [safe-sh](https://github.com/EpicEric/safe-sh) + static remainder; **1★**; + [jev-pastepilot](https://github.com/buberlo/jev-pastepilot) + Confirm before act; rh-guard owns); + **Retrieval as calibrated decision space** + ([Jev-Reranker](https://github.com/uspraveen/Jev-Reranker) + live Jev not yet measured; + [sessionwise](https://github.com/Nasrallah-AL/sessionwise) + opt-in relevance; + [jev-search](https://github.com/savka777/jev-search) + pointer sieve; **≠** kazuhideoki/jev-search + **≠** superagents-lab/jev-search); + **Enterprise reflexes** + ([400ms-agentic-sf](https://github.com/furuCRM-Inc/400ms-agentic-sf) + Salesforce WebMCP; + [typesafe-scheduler-diagnostics](https://github.com/thevilledev/typesafe-scheduler-diagnostics) + advisory); + **Screenshot-free / CU** + ([droidjev](https://github.com/mkruglikov/droidjev) + screenshot-free; + [jevcu](https://github.com/Tewoto1/Computer-use-and-control-with-Jev) + planner still writes); + **Hybrid S1/S2** + ([ha-conversation-jev](https://github.com/luxus/ha-conversation-jev) + Jev→Grok; **1★**; + [dsh-jev](https://github.com/buberlo/dsh-jev) + can only gate; **2★**); + **Harbor-jevals / SRE** + ([jev-classification-benchmark](https://github.com/rachit-srivastava-devx/jev-classification-benchmark) + specified not run; + [jev-luna-pagerduty-trigger](https://huggingface.co/datasets/reachjalil/jev-luna-pagerduty-trigger) + p≥0.50); + **Laya densifies** + ([meldecision](https://github.com/meldltd/meldecision) + laya-go ONNX; + [laya-doom](https://github.com/shantanugoel/laya-doom) + never pixels; + [laya-api](https://github.com/logixism/laya-api) + empty README; + [akpsahan/laya](https://huggingface.co/akpsahan/laya) + ≠ Archer); + **Demos / unofficial toolbelt** + ([jevchess](https://github.com/choxos/jevchess) + engine owns truth; **1★**; + [jev-drive](https://github.com/vedssharma/jev-drive) + sim not AV; + [story-arc](https://github.com/amali-s/story-arc) + Jev never authors; + [jev-hs-assistant](https://github.com/newbie1668/jev-hs-assistant) + HS6; + [jev-plays-starcraft-2](https://github.com/golergka/jev-plays-starcraft-2) + UI-verified ≠ API Victory; **1★**; + [awesome-jev-use-cases](https://github.com/walidboulanouar/awesome-jev-use-cases) + catalog; **2★**; + [typesafe-go](https://github.com/Nibir1/typesafe-go) + ≠ official). Soft Noul ≠ hard safety. 0★ HIGHs + still get real cards. Census (user-provided; not + re-derived): Archer still NOT landed; tracker likes + **50** lastModified UNCHANGED + 2026-09-19T18:37:18Z; SemIf 1873 (+7); jevlike 969 + (+2); TypeAR 10 flat; Awesomejev 561/27007 flat. + Do not copy `TYPESAFE_API_KEY` / OAuth `client_id` / + `uv` / `npx` / `.env`. No wrapper. No invented + metrics. +- SIGNAL fold jevcache + jev-align (`research/notes.md` + §93): **Skip Archer.** Docs-only on a **fresh PR + off main** after #9 merge `059f3670`. **Never + reopen** merged #3 / #4 / #5 / #7 / #8 / **#9**. + Do **not** re-fold 1441 / §92. How-to-apply / + mental models / architecture / Harbor-jevals / + toolbelt — not a thin Jev skill dump. + Backend-agnostic categorization/scoring/decision + class (Jev-like speed/econ). rh-guard owns + HIT-as-truth and training-score auto-accept as + gate cousins; Augustus owns placement. Formal + methods compose with scoring; a HIT and a + training score are SENSOR. Two clusters: + **Decision ledger / memoization** + ([hyperspaceai/jevcache](https://github.com/hyperspaceai/jevcache); + license **null**; **8★** this pass, SIGNAL ★6; + HEAD `a211d13`; README SHA `7c2abe99`; + fingerprint after redact; recall vs decide; + publish fingerprints+answers; CI replay as Harbor + cousin; Cache hit ≠ correctness; + hyperspaceai/jevcache ≠ kushals256/jevcache; + memoize typed decisions; VOI of cache hit); + **GEPA alignment loop** + ([sutro-sh/jev-align](https://github.com/sutro-sh/jev-align); + Apache-2.0; **60★** this pass, SIGNAL ★56; forks + **7**; HEAD `49753df`; README SHA `363fccb7`; + human labels only; score never auto-accepts; + production capture flywheel; sutro-sh/jev-align ≠ + caiovicentino/jev-align; GEPA + System One). Soft + Noul ≠ hard safety. Census (user-provided; not + re-derived): Archer still NOT landed. Do not copy + `TYPESAFE_API_KEY` / `uv` / `npx` / `curl | sh` / + `.env`. No wrapper. No invented metrics. +- SIGNAL fold enzyme + JA ModernBERT + Gemma/Nemotron/djev-dev/Laya + essay (`research/notes.md` §94): **Skip Archer.** Docs-only + on a **fresh PR off main** after #10 merge `30438eff`. + **Never reopen** merged #3 / #4 / #5 / #7 / #8 / #9 / + **#10**. Do **not** re-fold jevcache/jev-align / §93. + How-to-apply / mental models / architecture / + Harbor-jevals / toolbelt — not a thin Jev skill dump. + Backend-agnostic categorization/scoring/decision + class (Jev-like speed/econ). rh-guard owns + guidance-as-hook, unofficial-local-as-Jev, + hosted-bootstrap silent FALLBACK, + LFM-default-as-JA-softmax, Nemotron + “not calibrated replacement”, and Laya + confidence-without-competence as gate cousins; + Augustus owns placement. Formal methods compose + with scoring; `when asked`, unofficial local p, + Nemotron p, and Laya 0.85 are SENSOR. Three + clusters: **Compile-time System One / + questions-as-index** + ([byenzyme/enzyme](https://github.com/byenzyme/enzyme); + license **null**; **63★** this pass, SIGNAL ★62; + HEAD `c91d6b5`; README SHA `9af7c570`; + guidance ≠ hook; catalysts ≠ summaries; + compile-time System One; hosted bootstrap ≠ + silent TypeSafe; ~350×/1000× *theirs* + (GitHub description); **≠** enzymejs/enzyme); + **Unofficial JA ModernBERT cross-encoder** + ([argos1111/modernbert-ja-310m-jev](https://huggingface.co/argos1111/modernbert-ja-310m-jev); + CC-BY-SA-4.0; **2 likes**; sha `07cda235`; + unofficial ≠ TypeSafe; format_version + modernbert-jev/1; Argos1111/jev_local ≠ + us/jev-local ≠ kunchenguid/local-jev; LFM + default ≠ ModernBERT backend; JGLUE JNLI + 92.62% / JComQA 92.40% *theirs*); + **NAR class legitimacy / multimodal / + Router-OOD** + ([@googlegemma](https://x.com/googlegemma/status/2101069861598482817) + ~0.2s *theirs*; + [pst2154/Nemotron_Jev](https://github.com/pst2154/Nemotron_Jev) + **6★**; README SHA `f2f3d052`; HEAD `983cc29` on + `feat/nemotron-decision-lab`; Nemotron ≠ TypeSafe + Jev; not a calibrated replacement; + [Davipar/djev-dev](https://github.com/Davipar/djev-dev) + Apache-2.0; **2★**; README SHA `6d59d020`; + djev-dev complements djev-spark; images as Choice + options; + [Laya essay](https://laya.convaiinnovations.com/); + Laya essay numbers *theirs*; Router/OOD + confidence). Soft Noul ≠ hard safety. Census + (user-provided; not re-derived): Archer still NOT + landed. Do not copy `TYPESAFE_API_KEY` / + `ENZYME_JEV_MODEL` / `curl | bash` / docker / + `pip install laya`. No wrapper. No invented + metrics. +- Hourly 1541 HIGH (`research/notes.md` §95): **Skip + Archer.** Docs-only on a **fresh PR off main**. Never + reopen merged #7 / **#8** / **#9** / **#10** / + **#12**. Do **not** push onto §93 / §94. Do **not** re-fold + 1441 / §92. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class (Jev-like + speed/econ). rh-guard owns the injection-firewall / + CI-gate cousins; Augustus owns placement. Formal + methods compose with scoring; a Noul is a SENSOR; + pasting orchestrator 0.95 / skip_below 0.05 / + OpenRoboto $ as class ceilings, inventing + one-dollar-tahoe ASR/FPR, or treating numbered-choice + softmax as a Noul is soundness theater. Six clusters: + **Decision-as-plugin for SWE** + ([petercr/jev-orchestrator](https://github.com/petercr/jev-orchestrator) + difficulty + policy thresholds + JSONL trace; **0★**; + [Charlyhno-eng/jev-codex-pilot](https://github.com/Charlyhno-eng/jev-codex-pilot) + model + reasoning depth; **0★**; + [SunnyKikiHK/jev-replacement](https://github.com/SunnyKikiHK/jev-replacement) + keep/shadow/hybrid/reject; license null; **0★**); + **Evidence projection** + ([jackboykin/quarry](https://github.com/jackboykin/quarry) + quarry evidence projection; Go MIT; **0★**; **master**); + **Soft judgment integrity** + ([seb4ez/jevguard](https://github.com/seb4ez/jevguard) + calibrator/cache/escape; **0★**; + [guilhem/jev-ci-selector](https://github.com/guilhem/jev-ci-selector) + CI shadow mode; license null; **0★**; rh-guard owns); + **Physical/control first-class domain** + ([Frank-ZY-Dou/awesome-jev](https://github.com/Frank-ZY-Dou/awesome-jev) + Frank-ZY-Dou/awesome-jev robotics/3D/control; license + null; **0★**; text-state, not pixels); + **Harbor-jevals / injection-firewall** + ([PavitarSinghArneja/one-dollar-tahoe](https://github.com/PavitarSinghArneja/one-dollar-tahoe) + TypeSafe Jev defense eval; **0★**; ~74 demo; README + has no ASR/FPR; rh-guard owns); + **llama.cpp replica** + ([webNeat/llama-jev](https://github.com/webNeat/llama-jev) + llama.cpp replica; license null; **0★**; softmax ≠ + Noul). Soft Noul ≠ hard safety. 0★ HIGHs still get + real cards. Census **not provided this hour** (not + re-derived). Archer still NOT landed. Do not copy + `TYPESAFE_API_KEY` / `AI_GATEWAY_API_KEY` / + `EXA_API_KEY` / `GROQ_API_KEY` / `uv` / `npx` / + `go install` / `.env` / `attacks.json`. No wrapper. + No invented metrics. +- Hourly 1639 HIGH (`research/notes.md` §96): **Skip + Archer.** Docs-only on a **fresh PR off main**. Never + reopen merged #7–**#13**. Do **not** re-fold 1541 / + §95. How-to-apply / mental models / architecture / + Harbor-jevals / toolbelt — not a thin Jev skill dump. + Backend-agnostic categorization/scoring/decision + class. Formal methods compose with scoring; a Noul + is a SENSOR; treating zen-chat as calibrated Jev, + pasting tamaratran 24/24 onto this host, or + hard-gating keepThreshold 0.5 as proof of + irrelevance is soundness theater. One HIGH cluster: + **OpenCode host-port of evidence-preserving stdout + prune** + ([indiejoseph/opencode-jev-pruner](https://github.com/indiejoseph/opencode-jev-pruner) + OpenCode jev-pruner context sieve; + observe→score-candidates→prune; jev-zen / + jev-1.13-free; zen-chat ≠ Noul; fail-open original; + keepScore >0.1 floor; GitHub license **null**; + `package.json` MIT; **0★**; HEAD `764169c` + **master**). MEDIUM watch (do not over-weight): + [jongyunhur/jev-webagent-bench](https://github.com/jongyunhur/jev-webagent-bench) + empty stub; [Kiln-AI/jev_jsonschema](https://github.com/Kiln-AI/jev_jsonschema) + noul_threshold 0.5 (**5★**); [NSStudent/JevSwiftSDK](https://github.com/NSStudent/JevSwiftSDK) + unofficial (**5★**). Soft Noul ≠ hard safety. Census + **not provided this hour** (not re-derived). Archer + still NOT landed. Do not copy `TYPESAFE_API_KEY` / + `OPENCODE_API_KEY` / `npx` / plugin marketplace. + No wrapper. No invented metrics. +- SIGNAL gliner-native-runtime (`research/notes.md` §97): + **Skip Archer.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#14**. Do **not** re-fold 1639 + / §96. How-to-apply / mental models / architecture / + Harbor-jevals / toolbelt — not a thin Jev skill dump. + Backend-agnostic categorization/scoring/decision class. + User-linked SIGNAL (pushed 2026-08-26). Formal methods + compose with scoring; span confidence is a SENSOR; + treating this as TypeSafe Choice/Score/Noul, collapsing + it into Fastino official, or hard-gating default 0.1 as + NER quality is soundness theater. One HIGH cluster: + **GLiNER2 native Apple path** + ([shershah1024/gliner-native-runtime](https://github.com/shershah1024/gliner-native-runtime) + unofficial Swift/Core ML GLiNER 2.5-small; entity spans + + confidence; not Choice/Score/Noul; not TypeSafe; + label descriptions as schema; on-device ANE economics; + honesty locks; Apache-2.0; **4★**; HEAD `b44f661`; + README SHA `901eb063`). Soft Noul ≠ hard safety. + default threshold 0.1 still soft. Census **not provided + this hour** (not re-derived). Archer still NOT landed. + Do not copy `swift build` / Git LFS / Hub weights. + No wrapper. No invented metrics. +- Hourly 1740 HIGH (`research/notes.md` §98): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#15**. Do **not** re-fold 1639 + / §96 / gliner-native-runtime / §97 / 1541 / §95. How-to-apply + / mental models / architecture / Harbor-jevals / toolbelt — + not a thin Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; hard-gating DGP + as safety theater, hard-gating 0.4/0.2 as “concept absent,” + treating OpenRouter/TypeSafe auto-failover as one Noul + (silent FALLBACK), or pasting kev OOD 0.76 as + Jev-equivalent is soundness theater. Three HIGH clusters: **Decision Graph Protocol + envelope** + ([numerous-com/dgp](https://github.com/numerous-com/dgp) + Decision Graph Protocol frame→assess→commit; app retains + permissions/effects; Jev-first assessor-neutral; guarded + commit / receipt/next frame; assessment batching; + numerous-com/dgp ≠ TypeSafe official; Python MIT; **0★**; + HEAD `a9cb3c4`; README SHA `655fc638`); **calibrated + meaning-grep over a live tree** + ([can1357/jegrep](https://github.com/can1357/jegrep) + jegrep calibrated path+range Nouls; no + embeddings/index/daemon; ~$0.01–0.03 typical; agent --json; + can1357/jegrep ≠ Bentlybro/jevgrep ≠ uehaj/jev-semgrep; + Rust MIT; **13★**; HEAD `a280f14`; README SHA `6f241390`); + **Archer-arch fidelity + measured calibration gap** + ([jaredpalmer/kev](https://github.com/jaredpalmer/kev) + Archer-arch fidelity; kev family OOD 0.76–0.77 vs Jev 0.86; + block-causal isolation; pointer/readout CE-trained; + /v1/systemone drop-in; replica honesty; Apache-2.0; **507★**; + do not rewrite §45; Jev-omni owns the replica fold). Soft + Noul ≠ hard safety. 0★ HIGH still got a real card. Census + **not provided this hour** except Archer tracker likes + **51** (+1); lastModified UNCHANGED 2026-09-19T18:37:18Z; + Hub `archerhume/4rcherhume` HTTP **401**. Archer still NOT + landed. Do not copy keys / `uv` / `cargo` / `.env`. No + wrapper. No invented metrics. +- Hourly 2145 HIGH (`research/notes.md` §102): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#19**. Do **not** re-fold 2041 + / §101 / 1943 / §100 / 1843 / §99 / 1740 / §98 / 1639 + / §96 / gliner-native-runtime / §97 / 1541 / §95 / + jev-align *mechanism* / §93. How-to-apply / mental + models / architecture / Harbor-jevals / toolbelt — not + a thin Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating a + clean jevq run as measured separation, treating 62 + IC-Laya tests as Laya parity, letting a Score grant Tx, + pasting “Jev48 beats Jev on phishing” from AUROC, + injecting class priors as “help”, treating v4 as a + pharmacy controller, distilling Jev as teacher of + record, pasting maze 1.00 as a general System One, or + treating jev-kit exit 0 as claim truth is soundness + theater. Nine HIGH clusters: **question-linting of Jev + questions themselves** + ([yodablocks/jevq](https://github.com/yodablocks/jevq) + PRIMARY; nine jaggedness rules, no API key, no labelled + data; static lint ≠ measured separation; + yodablocks/jevq ≠ tenbin ≠ JevLint ≠ commitjev; Python + MIT; **0★**; HEAD `40b2dd90`; README SHA `3198dde0`); + **open-weights Laya as class exemplar (binding)** + ([ChristianAlexander/laya_ex](https://github.com/ChristianAlexander/laya_ex) + Nx/Bumblebee runtime; host chooses backend; Elixir + Apache-2.0; **0★**; HEAD `99f9ce73`; README SHA + `6361c920`; mix **0.1.0**); + **on-chain/edge Laya deploy** + ([humandebri/IC-Laya](https://github.com/humandebri/IC-Laya) + parity_verified stays false; model output never grants + Tx; Rust MIT; **0★**; HEAD `055ef42f`; README SHA + `85431606`); **auditable weekend replica** + ([agilabs-ai/jev48](https://github.com/agilabs-ai/jev48) + Jev outputs never used for training; unpaired 0.577 vs + 0.727; Python MIT; **0★**; HEAD `aa697005`; README SHA + `3f667dd4`); **adversarial dual-judge / framing** + ([copyleftdev/ember](https://github.com/copyleftdev/ember) + comparative framing is the usable judgment; prior + injection crowds out evidence; TypeScript MIT; **0★**; + HEAD `c02f622b`; README SHA `6db00b56`); **Laya + specialist + Hub replica** + ([PIXELZX0/XERON](https://github.com/PIXELZX0/XERON) + training still GPU-pending; license null; **0★**; HEAD + `5e870a4d`) + + ([daliborsb/laya](https://huggingface.co/daliborsb/laya) + Hub Laya replica drop; ≠ convaiinnovations/laya); + **distillation economics** + ([MagaBitmex/jev-4b-distill-data](https://huggingface.co/datasets/MagaBitmex/jev-4b-distill-data) + gold is programmatic; teacher is closed-API clone; do + not distill Jev as teacher of record; student checkpoint + missing); **non-LLM VIN System One** + ([lewislululu/jevon](https://huggingface.co/lewislululu/jevon) + planning depth not chat; maze 1.0000 n=141 *theirs*; + AGPL; likes **3**; ≠ douglance/jevon); **source-bound + evidence** + ([WaynezProg/jev-kit](https://github.com/WaynezProg/jev-kit) + local quote mismatch needs no API; exit 0 ≠ claim + truth; MIT; **0★**; HEAD `4558554f`; README SHA + `a7f14838`; ≠ jonathanavis96/jev-kit Airlock). Soft + Noul ≠ hard safety. 0★ HIGH still + got a real card. Census **live REST pulse** (review + relock): SemIf **2000★** (+16 vs §101 **1984**); + jevlike **1002★** **flat**; TypeAR **12★** (+1 vs + §101 **11**). Tracker likes **54**; + lastModified `2026-09-20T02:59:13Z`; Hub + `archerhume/4rcherhume` HTTP **401** (not re-fetched as + a rewrite). Archer still NOT landed. Awesomejev + 561/27007 user-provided (≠ AnotiaWang/awesome-jev + **86★**). Qwen3.8-27B ≠ Archer. + `invented_signal: false`. Do not copy keys / `npm` / + `pip` / `npx` / `uv` / `mix` / `curl | sh`. No wrapper. + No invented metrics. +- Hourly 2246 HIGH (`research/notes.md` §103): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#20**. Do **not** re-fold 2145 + / §102 / 2041 / §101 / 1943 / §100 / 1843 / §99 / 1740 + / §98 / 1639 / §96 / gliner-native-runtime / §97 / 1541 + / §95 / jev-align *mechanism* / §93. How-to-apply / + mental models / architecture / Harbor-jevals / toolbelt + — not a thin Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating a + catalog row as a bake-off win, treating constructed + scenes as production logs, pasting Hev 80.00% as Jev + identity, quoting mini-jev 93.25% as family-disjoint, + putting Jev in an AR next-token loop, laundering a Noul + as a Lean step, treating parallel rank-k as a sort + proof, pasting a curated list’s von sub-15ms as an + Augustus fact, or hard-thresholding paper-radar 0.5 as + frequency is soundness theater. ChatJev-style soundness + theater is the anti-pattern. Nine HIGH clusters: + **independent System One evidence catalog** + ([reachjalil/system-one-bench](https://github.com/reachjalil/system-one-bench) + PRIMARY; 19 reviewed records; scores not one leaderboard; no external record + currently reproduced; TokenTrim no-Jev matched hybrid + 62.4%; JavaScript MIT; **0★**; HEAD `ceb17269`; README + SHA `d9e0c7b7`; ≠ mallahyari/system-one-benchmark); + **typed eval freeze** + ([SivletLabs/jev-eval](https://github.com/SivletLabs/jev-eval) + 21 tasks · 134 items · 208 questions; scenes from public + GitHub contracts, not production logs; Python MIT; + **0★**; HEAD `3f9d976f`; README SHA `df16766c`; ≠ + willkelly/jev-evaluation ≠ 4esv/jev-eval); + **option-isolated tiny replica** + ([nafisazizir/hev](https://github.com/nafisazizir/hev) + option isolation (sibling-blind); permutation-equivariant; + Hub OWNER not published; Apache-2.0; **0★**; HEAD + `79a486f9`; README SHA `b995dce3`; ≠ jaredpalmer/kev); + **frozen-LLM typed decisions** + ([yuki-oshio/mini-jev](https://github.com/yuki-oshio/mini-jev) + frozen local LLM logits, no trained decision head; + residual-head 9,222-param decreased 73/96→67/96; + confidence = 1−normalized entropy, not P(correct); + Python MIT; **0★**; HEAD `dff5b323`; README SHA + `363441b6`; ≠ r-ms/mini-jev); + **AR next-token anti-pattern** + ([erik-dunteman/ChatJev](https://github.com/erik-dunteman/ChatJev) + Jev classifier as autoregressive next-token predictor; + ChatJev-style soundness theater; license null; **1★**; + HEAD `ea33ab8d`; README SHA `c763be19`; ≠ dannote/jev ≠ + jev-gpt); + **formal compose with scoring** + ([wufuju2023-cell/jev-alpha-proof-analysis](https://github.com/wufuju2023-cell/jev-alpha-proof-analysis) + calibrated decision head × AlphaProof value head; + implementation-layer isomorphism, semantic difference; + timeout = censoring; do not launder Noul as proof; + license null; **0★**; HEAD `afd9bb6f`; README SHA + `1810d7f6`); + **parallel rank vs serial selection** + ([zzzzzec/jevsort](https://github.com/zzzzzec/jevsort) + parallel rank-prediction vs serial selection; + independent questions can conflict; HTML; license null; + **1★**; HEAD `57067b90`; README SHA `85044740`; ≠ + keltokhy/jsort); + **open-side ecosystem catalog** + ([rupeshpoojary9/awesome-open-system-one](https://github.com/rupeshpoojary9/awesome-open-system-one) + curated open System One ecosystem catalog; CC0 1.0; + SPDX NOASSERTION; **0★**; HEAD `637ee3d3`; README SHA + `0007e343`; ≠ AnotiaWang/awesome-jev); + **knowledge-work paper radar** + ([LYchoon/paper-radar-jev](https://github.com/LYchoon/paper-radar-jev) + arXiv paper radar with Jev relevance scoring; ranking ≠ + calibration / 0.5 still soft; fail-open failed evals + not marked seen; Python MIT; **0★**; HEAD `fbadf01c`; + README SHA `1cb8a9c3`; size **73**; default master). + Soft Noul ≠ hard safety. 0★ HIGH still got a real card + (jevsort **1★**). Census **live REST pulse** (review + relock): SemIf **2019★** (+19 vs §102 **2000**); + jevlike **1006★** (+4 vs §102 **1002**); TypeAR **12★** + **flat**. Tracker likes **54**; lastModified + `2026-09-20T02:59:13Z`; Hub `archerhume/4rcherhume` + HTTP **401** (not re-fetched as a rewrite). Archer + still NOT landed. Awesomejev 561/27007 user-provided + (≠ AnotiaWang/awesome-jev **87★**). Qwen3.8-27B ≠ + Archer. `invented_signal: false`. Do not copy keys / + `npm` / `pip` / `npx` / `uv` / `.env`. No wrapper. + No invented metrics. +- Hourly 2041 HIGH (`research/notes.md` §101): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#18**. Do **not** re-fold 1943 + / §100 / 1843 / §99 / 1740 / §98 / 1639 / §96 / + gliner-native-runtime / §97 / 1541 / §95 / jev-align + *mechanism* / §93. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; treating a + zero binary name gap as a fairness certificate, letting + Jev emit the verdict string, pasting PrismNLI's lead + without the contamination caveat, treating J7 pass as + safe to obey, treating a receipt as authorization, or + hard-gating confidence ≥0.95 is soundness theater. Nine + HIGH clusters: **resume-screening bias audit** + ([natemoo-re/bias-bench](https://github.com/natemoo-re/bias-bench) + PRIMARY; name×resume factorial independent Nouls; + callback determined by resume quality; mean-probability + name gaps operationally negligible; natemoo-re/bias-bench + ≠ BBQ; JavaScript; license null; **0★**; HEAD + `fe2f2535`; README SHA `a1c3e604`); **MCDA panel + code-owned verdict** + ([austindixson/planalyzer](https://github.com/austindixson/planalyzer) + Plan/PRD panel → code-owned pass|review|block; + cheerleading out of scope; Python MIT; **0★**; HEAD + `39fc161f`; README SHA `6e4d8da3`); **EU cost-aware + routing** + ([cannacre8ive/switchboard-ai](https://github.com/cannacre8ive/switchboard-ai) + cost-aware multi-model routing/escalation; decide vs do; + successful-task cost; JavaScript MIT; **1★**; HEAD + `5cae9d1c`; README SHA `872de837`; package **0.4.0**); + **frozen-protocol class bake-off** + ([elcronos/jev-vs-open-decision-models](https://github.com/elcronos/jev-vs-open-decision-models) + TypeSafe Jev vs PrismNLI vs Laya; contamination caveat; + Python; license null; **0★**; HEAD `b61e6cfc`; README + SHA `b7256888`); **VOI admission** + ([cvsgireesh/jevusher](https://github.com/cvsgireesh/jevusher) + context-window admission control; fail polarity per + lens; on small inputs lenses lose money; TypeScript MIT; + **0★**; HEAD `d830d344`; README SHA `428a4a59`); + **Leveson control plane** + ([MokiMeow/jev-fabric](https://github.com/MokiMeow/jev-fabric) + typed decision control plane; receipt ≠ authorization; + historical-v0 zero retained cases; Apache-2.0; **0★**; + HEAD `95b9a4f3`; README SHA `f485dbdd`); **scoring + economics** + ([jose-troche/live-rubric](https://github.com/jose-troche/live-rubric) + live 15-dim typed rubric re-score per pause; OpenJev/Codiv + ≠ TypeSafe hosted; ~$0.000004 desc / ~$0.000006 README; + TypeScript; license null; **0★**; HEAD `db8da8db`; + README SHA `4a0be084`); **pre-registered calibration + science** + ([willkelly/jev-evaluation](https://github.com/willkelly/jev-evaluation) + 28 predictions before data; 123,805 requests; confidence + does not track ignorance; polite injection 65% / crude + 0%; Python MIT; **0★**; HEAD `c168c093`; README SHA + `2d66ac22`; rh-guard owns injection); **class + infrastructure SDK** + ([nshkrdotcom/system_one_sdk](https://github.com/nshkrdotcom/system_one_sdk) + provider-neutral Elixir/BEAM Noul/Choice/Score SDK; + GitHub desc provider-neutral / README TypeSafe-first; + Elixir MIT; **0★**; HEAD `c2a522ee`; README SHA + `c117b4c4`; mix **0.5.0**). Soft Noul ≠ hard safety. + 0★ HIGH still got a real card (switchboard now **1★** + live REST). Census **live REST pulse**: SemIf **1984★**; + jevlike **1002★**; TypeAR **11★** flat. Tracker likes + **54** (+3 vs §100 pin **51**); lastModified **CHANGED** + 2026-09-20T02:59:13Z; Hub `archerhume/4rcherhume` HTTP + **401** (not re-fetched as a rewrite). Archer still NOT + landed. Awesomejev 561/27007 user-provided (≠ + AnotiaWang/awesome-jev 84★). + `invented_signal: false`. Do not copy keys / `npm` / + `pip` / `npx` / `uv` / `mix`. No wrapper. No invented + metrics. +- Hourly 1943 HIGH (`research/notes.md` §100): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#17**. Do **not** re-fold 1843 + / §99 / 1740 / §98 / 1639 / §96 / gliner-native-runtime + / §97 / 1541 / §95 / jev-align *mechanism* / §93. + How-to-apply / mental models / architecture / + Harbor-jevals / toolbelt — not a thin Jev skill dump. + Backend-agnostic categorization/scoring/decision class. + Formal methods compose with scoring; a Noul is a SENSOR; + hard-gating `feels` 80%, pasting 17/18 as Harbor, + hard-gating 0 of 157, treating boolean @ 0.5 as a proof, + claiming 10×, treating softmax as a Noul, or treating 62 + tests as quality is soundness theater. Six HIGH clusters: + **Jev IS the if-statement** + ([southpolesteve/probably](https://github.com/southpolesteve/probably) + PRIMARY; judgments/probabilities drive branches; text + model only writes prose; interpreter owns + variables/loops/budgets/replay; otherwise maybe / + confidence gate; chaos samples after the gate; + southpolesteve/probably ≠ carldaws/hunch ≠ feelings ≠ + Kungie/gut ≠ Illusion47586/judge ≠ tidymodels/probably; + TypeScript MIT; **3★**; HEAD `6bf671a4`; README SHA + `c28570a9`); **GEPA live delta** + ([sutro-sh/jev-align](https://github.com/sutro-sh/jev-align) + 133★ / forks 10 live; build calibrated classifiers from + human feedback; HEAD `49753df9` / README SHA `363fccb7` + unchanged vs §93); **retrieve by relevance not + resemblance** + ([samdotmak/jev-recall](https://github.com/samdotmak/jev-recall) + one calibrated yes/no per memory in one request; pointer + mode 17/18 19/20 *theirs*; embedding resemblance misses + the allergy; MIT; **6★**; HEAD `d3e4acfb`; README SHA + `ea774519`); **memory leases ended by new evidence** + ([chopratejas/invalidate](https://github.com/chopratejas/invalidate) + HIGH upgrade; six Nouls then fixed rules in code; 0 of + 157 false invalidations; questions/plans/directives are + not evidence; unsure → review queue; host keeps the + store; Apache-2.0; **11★**; HEAD `d6ade601`; README SHA + `8cccad5f`); **contract-of-artifact lint rename** + ([mizchi/jev-lint](https://github.com/mizchi/jev-lint) + is mizchi/jevlint rename; name↔body / comment truth / + test-claims; no shipped rule has severity error; ~1 in 5 + findings wrong *theirs*; TypeScript MIT; **13★**; HEAD + `62d73f8e`; README SHA `4c0e37cd`); **JSON Schema + question compiler** + ([Kiln-AI/jev_jsonschema](https://github.com/Kiln-AI/jev_jsonschema) + HIGH upgrade; JSON Schema → typed JSON via Jev; + noul_threshold 0.5 decoder not a proof; + IncompatibleSchemaError lists every bad property; MIT; + **5★**; HEAD `fccea8c2` unchanged); **local System One + economics** + ([mizorewww/laya-coreml](https://github.com/mizorewww/laya-coreml) + on-device Laya CoreML ANE; ~5 ms P50 short decisions; + 189/189 FP16 checkpoint parity; 10× not achieved; + Apache-2.0; **0★**; HEAD `47f4baf0`; README SHA + `2068c661`) + + ([Micha0827/snapjudge](https://github.com/Micha0827/snapjudge) + softmax over allowed tokens ≠ Noul; question-first + cache; Python MIT; **3★**; HEAD `2df5ce27`; README SHA + `64a91458`) + + ([direwolfiy/JevPi](https://github.com/direwolfiy/JevPi) + Jev-first Pi agent loop; slow-LLM fallback; explicit + action menu / CandidateSource unimplemented; 62 tests + wiring not quality; license null; **0★**; HEAD + `980f8895`; README SHA `88ec1a55`). Soft Noul ≠ hard + safety. 0★ HIGH still got a real card. Census **live + REST pulse**: SemIf **1954★**; jevlike **989★**; TypeAR + **11★** flat. Tracker likes **51** flat; lastModified + UNCHANGED 2026-09-19T18:37:18Z; Hub + `archerhume/4rcherhume` HTTP **401** (not re-fetched as + a rewrite). Archer still NOT landed. Awesomejev + 561/27007 user-provided (≠ AnotiaWang/awesome-jev 83★). + `invented_signal: false`. Do not copy keys / `bun` / + `pip` / `npx` / `uv` / HF download. No wrapper. No + invented metrics. +- Hourly 1843 HIGH (`research/notes.md` §99): **Skip + Archer rewrite.** Docs-only on a **fresh PR off main**. + Never reopen merged #7–**#16**. Do **not** re-fold 1740 + / §98 / 1639 / §96 / gliner-native-runtime / §97 / + 1541 / §95. How-to-apply / mental models / + architecture / Harbor-jevals / toolbelt — not a thin + Jev skill dump. Backend-agnostic + categorization/scoring/decision class. Formal methods + compose with scoring; a Noul is a SENSOR; hard-gating + cost-derived 0.038 as a proof, merging von Needle + 52.6% with n=78 93%, “guaranteeing” calibration, or + pasting “Jev wins guardrailing” is soundness theater. + Five HIGH clusters: **cost-derived YES/NO/UNSURE + control flow** + ([Kungie/gut](https://github.com/Kungie/gut) + PRIMARY; cost-sensitive decision theory × System One + probabilities → control flow; thresholds derived from + costs not hard-coded; YES / NO / UNSURE from + cost_false_yes / cost_false_no / cost_human; + auto-batching same-object questions; Kungie/gut ≠ + tpellet/hunch ≠ carldaws/hunch; GitHub Apache-2.0 / + LICENSE MIT / pyproject Apache-2.0 *theirs*; **0★**; + HEAD `cb56c875`; README SHA `630474f6`; pre-alpha); + **typed-callback twin** + ([Illusion47586/judge](https://github.com/Illusion47586/judge) + judgment vs generation; deterministic execution after + probabilistic judgment; exactly one app-owned + callback; explicit uncertain branch; + Illusion47586/judge ≠ lexingtonhibiki/judgekit ≠ + Ascurse/typed-judge-kit; TypeScript MIT; **0★**; HEAD + `e69f65a1`; README SHA `08554c6f`; `@brkn-labs/judge` + 0.1.0); **variable-N option scoring as the trainable + object** + ([zwliJay/jev-forge](https://github.com/zwliJay/jev-forge) + dynamic candidate bags not fixed label sets; + zwliJay/jev-forge ≠ NanoJev; class-architecture note, + not a sixth species; GitHub NOASSERTION / LICENSE MIT; + **1★**; HEAD `eb3e4a2d`; README SHA `359f3f57`; do not + clone Hub weights); **open NAR replica economics** + ([wfzyx/von](https://github.com/wfzyx/von) late-catch + HIGH of the same repo; NAR local drop-in; open replica + economics / latency vs closed Jev; competing NAR + claims / replica honesty; Apache-2.0; **43★**; HEAD + `b9e42b26`; README SHA `574aa628`; do not merge Needle + 52.6% with n=78 93.0%); **typed vs chat judges on + guardrailing** + ([ishaannk/llm-vs-jev](https://github.com/ishaannk/llm-vs-jev) + cross-note only; nothing wins outright; can be argued + out of guarding; deeper integrity fold is rh-guard; + Apache-2.0; **0★**; HEAD `182e0864`; README SHA + `d9ebd40f`). gut/judge are **control-flow / + decision-theory overlays, not new class-table + species**. Soft Noul ≠ hard safety. 0★ HIGH still got + a real card. Census **live REST quoted**: SemIf + **1936★**; jevlike **983★** (watch claimed 984); + TypeAR **11★** flat. Tracker likes **51** flat; + lastModified UNCHANGED 2026-09-19T18:37:18Z; Hub + `archerhume/4rcherhume` HTTP **401**. Archer still NOT + landed. Awesomejev 561/27007 user-provided (≠ + AnotiaWang/awesome-jev 83★). X MCP `since_id` held; + pages_archived 0; no invented tweets. + `invented_signal: false`. Do not copy keys / `pip` / + `npm` / `uv` / Hub download. No wrapper. No invented + metrics. +- Effect-oriented loops (`notes.md` §28, `mappings.md` §19): Ward's + ZIO client keeps Jev as the outer Choice and the handler as the + effect. Not Effect.ts. GLiNER author: GLiNER2 "like jev" is GLiGuard + schema-conditioned categorize, not a Noul. +- Boundary-audit stop conditions for TOCTOU-of-Noul and vacuous specs; + FAQ rows for Alloy vs Apalache and PufferLib-as-DST-trio +- Research pointer to [dayhaysoos/jevals](https://github.com/dayhaysoos/jevals): + local MIT workbench for Jev questions vs labeled Noul/Choice/Score cases + (compare runs, WebMCP + agent skill). Empirical acceptance-test surface + for Hypothesis mapping cards; complements `evaluate_decisions.py`. Not a + jevals how-to (`research/notes.md` §24; one sentence in `validation.md`) +- Mental-models card: Augustus is design judgment across AI, SWE, + business, knowledge work, and life — not SWE-only. Pillars: expected + utility / selective classification, calibration and cost-sensitive + thresholds, VOI, MCDA, search/control substitutions, signal detection, + Leveson org/safety, NATM/snap-fit/Norman/Kent/Shirky as general + intuition. Domain gallery labeled Hypothesis except launch-week + Empirical SWE rows. +- Archer Hume architecture reconstruction (17 Sep 2026 essay, ~10k + probes of `jev-1.13.0`): direct readout vs generated confidence, + isolated questions, listwise IIA and order sensitivity, confidence as + arithmetic on the distribution. Independent envelope probe; does not + override live TypeSafe docs. Announced open-weight drop is **WATCH** + (27B dense, AU healthcare residency, prefers "decision models"; still + no Hub weights). `research/notes.md` §31, §33; `judgment-class.md` + when-to-use table; FAQ confidence / surfaces questions. +- Entropy as allocator (**Hypothesis**, `judgment-class.md`): Atallah's + low / medium / high buckets place System One on typed decisions and a + frontier decoder on high-entropy synthesis — same axis as marginals + vs joint and as VOI. "Review this PR" as medium is still partly + generative; "first model ever" is a claim. `research/notes.md` §38 +- Marginals, not a probabilistic program (`judgment-class.md`, FAQ): + Erik Meijer — Jev is a cool API and not a PPL; Kleisli qualifications + exaggerate; "Jev gives you the marginals; a decoder gives you the + joint." Joints and invariants stay with TLA+ / Alloy / contracts. + `research/notes.md` §34 +- Bespoke Nimble: open contrastive recipe, not a Jev distill. Model + card Apache-2.0 LoRA on Qwen3.5-9B (repo license absent). Their + 324-example holdout is a named receipt (Nimble 90.12%, Jev 1.13.0 + 93.21%), not a ranking. 9B-vs-Jev on your labels stays Hypothesis. + `research/notes.md` §35; one sentence in `validation.md` +- djev-spark: third compute graph (diffusion structured reads, + Jev-shaped I/O, image-in). Empirical as the public interface; + Hypothesis that it beats a decision head on your task. Archer's + multimodal drop stays WATCH. `research/notes.md` §36 +- Perception specialist then judgment specialist vs shared multimodal + System One (**Hypothesis**): SAM 3.1 (masks and tracks) or an ASR + transcript, then typed decisions on that state, is an application + pattern, not native omni. Information dies at the interface. Prefer + a shared multimodal decision model when the joint matters (Archer + Watch, not Empirical; djev-spark images; future audio). Basit ask, + primary post not retrieved. `research/notes.md` §39 +- Perception→decision pipeline, measure, and hill-climb + (**Hypothesis**, `validation.md`): stages with a versioned state + contract; stage metrics plus a frozen taskset; HoH changes one stage + or one interface. DSPy/Ax only on LM-program knobs; jevals and + calibration for the decision slice; Harbor names product + end-to-end, not a tutorial. `research/notes.md` §41 +- Eval & hill-climb (`validation.md`): jevals decision-stage hygiene + (independent keys, correctness is not confidence, held-out, immutable + runs) and Harbor as the product taskset substrate; one composition + table. `research/notes.md` §40 + +### Changed + +- Skill description rewritten as trigger conditions (mixed architecture, + prefilter, routing, preference lint, classification skepticism, family + choice including GLiNER/GLiClass/listwise/vision) plus an explicit `not_for` + against the official `typesafe-ai` skill +- Identity lock vs neighbor skills (`typesafe-ai`, `tenbin`, `decision-first`) + so Augustus stays the design-judgment layer — class-wide, not TypeSafe-only +- Design cards name hole, family, and typed judgment provider (Jev default; + other family only with self-eval) +- Protocol fan-out step is family-aware (Jev batch, GLiClass one-pass, + dual-encoder prompt scoring); ranking vs decision fail policy is a + non-negotiable +- Protocol and FAQ branch for "formally verify with Jev"; methods-catalog + and composition-algebra verifier position point at the ownership split +- Skill mission and description are domain-general (AI / SWE / business / + knowledge work / life); FAQ "is this only for software?"; mappings.md + beyond-SWE examples labeled Hypothesis; boundary-audit red flags for + TOCTOU-of-Noul and vacuous specs; formal-methods expanded with Alloy vs + Apalache and the DST trio including PufferLib; GLiNER promoted from + cousin footnote to species-map peer + +### Fixed + +Adversarial review of the whole skill against its own non-negotiables +(findings in `research/notes.md` §27). + +- Gate fail policy is per action, not universally open + (`composition-algebra.md` position 3, `agent-self-assessment.md`): + advisory guards fail open *because* an interlock sits underneath; + selection and authorization gates fail closed +- Dual-orchestration topology A selects from a closed catalog instead of + "planning" MCP calls, which contradicted the standing planner rejection +- Species map applied to the skill's own advice: GLiClass (categorize) is + the large-catalog substitute for a 255-option Choice; GLiNER spans are + not (`SKILL.md`, `judgment-class.md`, `applied-mappings.md`) +- Han Xiao trolley relabeled an Empirical **rejection** (one tweet, no + repo), not a recipe +- openjev-lm caveat moved to the figure it belongs to: 92.9% is against 70 + hand-labelled gold, 98.1% is teacher *agreement* +- Contract surface removed from design cards: the Ax constructor call and + the `instructions` key enumeration point at live docs instead + (`optimizer-integration.md`, `question-design.md`) +- `mappings.md` preamble no longer claims uniform Hypothesis where card + bodies say Contract/Empirical; §17 forbids reusing jevgate's ≤0.18 as a + constant; all Hypothesis-range references aligned to §6–§19 +- Ownership split labeled Contract in `toolbox-mapping.md`, matching + `mappings.md` §8; done-check splits structure from the Noul + +Second pass on `7b3a0c3` (`research/notes.md` §43). Zero blockers. +Dropped the unpublished `npx jevals` line; SAM and ASR are upstream +producers, not the perceive species; removed two call shapes from +`optimizer-integration.md`; tagged the $0.042/MTok cell as a vendor +figure; marked GodsBoy 94.4% exploratory. +- Skill description gained trigger terms for boundary audit, question + diagnosis, agent self-supervision, and optimizer placement + +## [0.2.0] - 2026-09-18 + +Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) +(`65a39f3`, 2026-09-12). + +### Added + +- Boundary-audit card for existing systems: three-way split (exact / + bounded judgment / generation), code-smell catalog, fit test, opportunity + map, smallest-viable-boundary rule, Jev-around-LLM sandwich, centralized + policy + raw-judgment retention, red flags, completion questions +- Protocol branch: audit a codebase/PR before inventing mappings; per-action + risk gates; keep questions/thresholds in one reviewable module +- Skill description trigger terms for brittle parsers, prompt-to-JSON + classifiers, and agent loops that are really bounded decisions + +## [0.1.0] - 2026-09-18 + +Pegged against [`typesafe-ai/skills` v0.5.7](https://github.com/typesafe-ai/skills/tree/v0.5.7) +(`65a39f3`, 2026-09-12) — the only tagged revision of the official skill at +Augustus launch. + +### Added + +- Skill protocol, decision-design card, and evidence labels (Contract / + Empirical recipe / Hypothesis) +- Classical-method mappings: features/utility, selective decisions, decision + circuits, bounded rerank, hierarchy/beam search +- Agent self-assessment, optimizer coupling (Ax, DSPy, ProgramAsWeights) +- Toolbox sweep, named-methods catalog, 11-position composition algebra, + question-design diagnosis +- Validation gates and `scripts/evaluate_decisions.py` +- Launch-week evidence archive (187 repos) and public ecosystem index +- Claude Code marketplace manifest; public GitHub mirror at + [`24601/Augustus`](https://github.com/24601/Augustus) + +### Research log (pre-tag) + +The dated passes below are how 0.1.0 was assembled. + +### 2026-09-18 (refresh pass 2) +- Research: Gemini Deep Research report retrieved and archived (interaction ID + saved in jev-archive/state); GitHub census doubled to ~60 Jev repos + + framework integrations (LiteLLM, LangChain, Vercel AI, Mastra, Eliza, Ax, + Composio); all 87 repos cloned to /home/user/workspace/jev-archive for + hourly refresh. +- New measured recipes added to notes.md: foreman supervision loop, pi-jev + gate thresholds, pi-warden 6→0 paired-run result, winnow relevance sieve, + fast-jev-compaction two-noul rule, skill-router gates (0.30/0.40, shortlist + 3, 94.4% vs 70.8%), calibration ECE 0.0313 vs 32% OOD collapse (Archer + Hume), Every 777-judgment eval, Near Here moderation numbers. +- Skill: added references/agent-self-assessment.md (agent self-supervision + lifecycle, grounding/citation checks, skill callability testing) and two + mapping-index rows; validation.md dogfooding section still canonical. + +### 2026-09-17/18 (initial) +- Baseline research archive (sources.json, notes.md), augustus skill with + mappings + validation references, evaluator script, hourly refresh script, + Claude plugin marketplace manifest. + +### 2026-09-18 (topic-index pass 3) +- Fixed census method: exact GitHub search paginated (700 repos created since + 09-14 captured; 700-result cap noted) + topics/jev crawl → ~80 additional + repos; archive now 184 clones. Miss-cause documented: earlier star-sorted + limit-40 search cut the low-star tail (incl. both MCTS repos). +- Skill: MCTS mapping promoted experimental → empirical recipe (grounded vs + speculative fidelity in types; probes-only concession; measured 24/24 vs + 1/24 greedy); agent-self-assessment.md gains the judge-variance recipe + (Jev judge 224-279x more consistent than LLM judge over 100 reps). + +### 2026-09-18 (pass 4 — optimizers + official skills + clone audit) +- ax Jev support documented from source (native adapter details, trueThreshold + semantics, fail-closed mapping validation); new reference + optimizer-integration.md covering Ax + DSPy typesafeify + jev-dspy-lab. +- typesafeainate/dspy-typesafeify cloned; official typesafe-ai/skills already + archived and layered-on (never duplicated). +- Clone audit: repos.txt deduped (185 unique), 0 missing on disk, no failures. + +### 2026-09-18 (pass 5 — toolbox sweep meta-method) +- New references/toolbox-mapping.md: the how-to-find-approaches-and- + applications procedure (judgment-shaped-hole substitution, newly-feasible + classification via economics inversion, standing rejections list); wired + into SKILL.md central model + index row. + +### 2026-09-18 (pass 6 — named-methods + operators/theorems tier) +- references/methods-catalog.md: ~20 named algorithms (CatBoost row is + Empirical via autoresearch cookbook) + operators/theorems tier with + precondition-carrying rule; wired into SKILL.md index and toolbox sweep. + +### 2026-09-18 (pass 7 — composition algebra as application generator) +- references/composition-algebra.md: 11-position grammar of Jev-vs-construct + relations, logical-operator combination rules, and the position×construct + traversal as the systematic application generator; wired into SKILL.md + index + toolbox sweep.