feat: repo contract pilot - distilled AGENTS.md, verify skill, REVIEW.md, cloud wiring - #470
Merged
Conversation
….md symlink), REVIEW.md, verify-temperpaw skill + feature map, cloud global-context hook, stack plugin declaration - AGENTS.md: global duplications removed (worktree/PR/root-cause/e2e boilerplate now comes from the stack layer every harness loads); project substance kept and tightened; new WASM integration rules (integrations never dispatch transitions; one integration, one concern) - CLAUDE.md -> symlink to AGENTS.md (one file, no hand-mirroring) - REVIEW.md: four passes incl. entity-first compliance and determinism - .claude/skills/verify-temperpaw/ + features/ (generated via create-verification-skill; .cursor/skills symlink for Cursor) - .claude/hooks/global-context.sh + settings.json: cloud sessions get the global layer and the stack plugin; local sessions unaffected (no duplication) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VLPhB9kjLeE48kLUyAXXq2
…d from proof run - Makefile wasm target now also runs os-apps/*/wasm/*/build.sh (paw-fs keeps per-module scripts one level deeper; the shallow glob left blob_adapter, workspace_fs, artifact_batch_apply unbuilt and the server shut down on missing AppRequired modules) - verify-temperpaw skill: TEMPER_API_KEY must be in .env BEFORE boot (the platform bootstraps the tenant credential at startup; keyless boot serves 401 on /tdata), service-doc-401 gotcha, isolation via TURSO_URL scratch DB (.scratch/ gitignored) Proven end to end: make wasm (12 modules) -> boot -> healthz 200 -> authed OData read 200 on /tdata/SkillPackages -> clean teardown. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VLPhB9kjLeE48kLUyAXXq2
rita-aga
force-pushed
the
claude/repo-contract
branch
from
August 25, 2026 23:28
4f23ff5 to
a6ac363
Compare
…fo/exclude '.claude/skills/' - added with -f) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VLPhB9kjLeE48kLUyAXXq2
Rita is rethinking how work gets proven; the old regime goes now, the replacement is an open design item. - .proofs/ (166 files), .progress/, docs/proofs/ deleted - 25 scripts deleted: prove_*, one-off seeds/calibration - none referenced by CI, Makefile, Dockerfile, or tests (each cross-checked) - tests un-enforce the regime: .proofs gitignore/dockerignore assertions and path exemptions removed (native_skill_installation, identity contract) - AGENTS.md: evidence-in-.proofs line removed - docs/datadog-webhook-setup.md: references to deleted scripts removed Kept: prove_wasm_guest_observability.py, read_datadog_snapshot.py, configure_metric_percentiles.py (bound by the Datadog observability contract test - untangling that belongs to the new proof design). Contract tests pass: native_skill_installation 5/5, identity contract 12/12. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VLPhB9kjLeE48kLUyAXXq2
…t definitions .claude/skills, .claude/agents, .cursor/skills are now symlinks into .agents/. Content exists once; harness folders only point. Also sidesteps the local .git/info/exclude rule that silently untracked .claude/skills. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VLPhB9kjLeE48kLUyAXXq2
rita-aga
marked this pull request as ready for review
August 26, 2026 16:36
|
Too many files changed for review (222 files, 100 file limit). Bypass the limit by tagging |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Why
Pilot of the universal-setup repo contract (stack ARN-405, Stage-1 item 2). TemperPaw first.
Scope
Decisions & Tradeoffs
Verification
Proof run complete (2026-08-25, local, isolated scratch DB). The skill's own instructions were executed end to end and each failure became a fix in this PR:
make setupnever builds WASM, and the oldmake wasmtarget was broken for the whole workspace. Fixed:make wasmnow runs everyos-apps/*/wasm/build.shand per-moduleos-apps/*/wasm/*/build.sh(paw-fs keeps its scripts one level deeper; missing them leftblob_adapter,workspace_fs,artifact_batch_applyunbuilt and the server shut down on missing AppRequired modules)./tdataauth:TEMPER_API_KEYmust be in.envBEFORE boot (platform bootstraps the tenant credential at startup; keyless boot serves 401). Skill Launch corrected.~/.local/share/temperpaw/paw.db- the skill now mandatesTURSO_URL=file:.scratch/paw.db(gitignored).Final pass:
make wasm(12 modules, all built) ->cargo run -p temperpaw->GET /healthz200 -> authed OData readGET /tdata/SkillPackages200 -> clean teardown by captured PID. Evidence:/tmp/verify-temperpaw/2026-08-25/(healthz response, boot log heads, OData payload).Authored by Fable 5 via Claude Code.
🤖 Generated with Claude Code
https://claude.ai/code/session_01VLPhB9kjLeE48kLUyAXXq2