docs(site): npm 0.6.2 landing copy and readiness tests

[dancingteeth]

Site-only bump for @dancingteeth/agent-looper 0.6.2 (2026-09-21), ready for merge + rsync.

Changes

• Version — 0.6.2 in site/llms.txt, changelog links on about/docs/index; supported line stays 0.6.x; no 0.6.1 current-npm leftovers.
• Cursor + presets — Document that runtime: cursor does not inherit repo OpenCode Go escalateModel; judge path is reviewRuntime / reviewModel (homepage FAQ + JSON-LD, harnesses Cursor card).
• agent-loop-prompt — Scaffold judge stops when frozen GOAL.md + verify.sh pass freeze (10-minute wall) on homepage FAQ + llms.txt.
• DSH credentials — Harness card note: version: "1" (quoted); bare YAML integer breaks headless boot.

Verify

• pnpm exec vitest run src/site/landingAgentReadiness.test.ts — 29 passed

  

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: 27329117-2662-43e6-8644-49ce2a428b5a

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[dancingteeth]

Reviewey Ray sensor — not a merge decision. Human owns merge.

dancingteeth/agent-looper PR #36 · head 2eb213ee · counter=26 · short sensor (Pass 0+1)

docs(site): npm 0.6.2 landing copy and readiness tests — Cursor escalateModel caveat, scaffold 10-minute wall, DSH credentials version: "1"; bump changelog/llms labels; extend landingAgentReadiness.test.ts. Delta 4c76784b→2eb213ee (+43/−20, 10 files). Agent-authored (cursoragent). PR already merged (3ffe8a05) at review time.

Risk

LOW

What could go wrong?

• Site / llms.txt "Current npm" and changelog chips say 0.6.2 while package.json and published npm are 0.6.3 — agents and humans may install or cite the wrong line.
• DSH note only says quote version: "1"; CHANGELOG 0.6.3 adds that wrapped credentials stores still abort (quoting alone is not enough).
• Copy claims (Cursor escalate drop, scaffold wall) not re-verified hands-on in this short sensor; readiness tests cover string presence only.

Review depth

• Journeys at risk: none (docs/site + readiness fixtures; no harness runtime src)
• Line-by-line: site HTML/MD/llms.txt version + harness notes; readiness test expectations
• Empirical checks: compared package.json (0.6.3), npm view (0.6.3), CHANGELOG 0.6.2/0.6.3 headlines vs site labels; CI test (22)/(24) still in progress on merge
• Reversibility: easy (docs-only revert / follow-up bump)
• Verification gap: adjacent-only (string readiness tests; no live Cursor/DSH smoke)
• Release guardrails: REVIEWS.md Loop risk → LOW documentation-only; §10 skip full pincer on short sensor
• Change set: base 4c76784b → head 2eb213ee

Overlay (Pass 1b)

REVIEWS.md: LOW = "documentation-only"; sensor contract PASS | ADVISORY | BLOCKERS; non-empty Blockers ⇒ BLOCKERS.

Verdict

ADVISORY

Advisory

• Stale Current npm pin — landing/llms.txt/about/docs changelog chips were moved 0.6.3 → 0.6.2 while package + npm stay on 0.6.3. Prefer a follow-up that restores 0.6.3 (or documents why the site deliberately lags).
• DSH credentials guidance lag — PR documents quoted version: "1" (0.6.2 headline); 0.6.3 changelog says wrapped version/refs/records stores still fail before WORKER. Worth aligning harness card copy with 0.6.3.
• Already merged — sensor is retrospective; CI Node 22/24 was still in progress at review. Not a merge gate.

Nits

• Test title/body still center "0.6.2" after main had already moved readiness expectations to 0.6.3 — expect churn on the next version bump.

[dancingteeth]

Reviewey Ray sensor (not a merge decision)

Risk

LOW

What could go wrong?

• Site/llms.txt now say Current npm: 0.6.2 and changelog chips point at 0.6.2, while registry npm view @dancingteeth/agent-looper version is 0.6.3 and CHANGELOG.md on main already documents 0.6.3 — 2026-09-21.
• DSH harness note teaches version: "1" (quoted) as the fix; CHANGELOG 0.6.3 says quoting is not enough for a wrapped credentials store (version / refs / records) — operators may follow stale site copy after rsync.
• Cursor escalate / scaffold 10-minute wall / readiness-test extensions look consistent with CHANGELOG 0.6.2 headlines; version pin is the mismatch.

Review depth

• Journeys at risk: none (site + readiness tests only; no harness runtime)
• Line-by-line: site/** copy + src/site/landingAgentReadiness.test.ts assertions for 0.6.2 / Cursor escalate / scaffold wall / DSH version: "1"
• Empirical checks: npm view → 0.6.3; CHANGELOG head has 0.6.3 then 0.6.2; CI test (22) / test (24) in progress at review; CodeRabbit SUCCESS
• Reversibility: easy (docs/site)
• Verification gap: adjacent-only (PR body cites 29 vitest pass locally; CI not green yet here)
• Release guardrails: PR already MERGED (3ffe8a05); human owns post-merge site rsync / version fix
• Change set: base 4c76784b → head 2eb213ee (PR docs(site): npm 0.6.2 landing copy and readiness tests #36; merge 3ffe8a05)
• Cadence: counter=26 short sensor (REVIEWS.md documentation-only → LOW; not a 10th)
• Overlay: REVIEWS.md Loop risk LOW = documentation-only

Verdict

ADVISORY

Advisory

• Align site/llms.txt/changelog chips with published npm (0.6.3 today) or publish/explain why the landing still pins 0.6.2.
• Refresh DSH credentials copy against CHANGELOG 0.6.3 (flat map; wrapped store still fails; quoting alone is insufficient).
• Wait for CI test (22) / test (24) if not already green before treating readiness tests as verified on this head.

Not a merge. Human owns merge / post-merge rsync.