Repository navigation
docs(agents): list every workflow in the CI lane map - #1431
Conversation
The CI lane map table was written when main carried 12 workflow files and never picked up two that landed later: ci-kani.yml (Kani bounded-verification harnesses over the rcache/sessions security cores, added for #1109) and spec-lint.yml (the advisory spec-format linter installed from gominimal/foundry). Add both rows, grouped with their respective ci-* and PR-gate neighbors, and bump the workflow count from 12 to 14. Also note that ci-kani-success is not part of the five required lane aggregators: ci-kani.yml's own header says the lane stays advisory until its harnesses have soaked, and is promoted by adding ci-kani-success to the branch-protection ruleset with no workflow change needed. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Essentials Run ID: 📒 Files selected for processing (1)
Included review availability: Your plan provides up to 5 included reviews per hour; 0 remain after this review. 📝 WalkthroughWalkthrough
ChangesCI documentation
Priority: ➖ Normal Estimated code review effort: 1 (Trivial) | ~5 minutes Change: Other · Severity of issue fixed: Medium Suggested reviewers: Merge Risk: ⚪ Minimal · up to This documentation-only change has no identified unresolved production or merge risk. 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
A rabbit checks the lanes at night Comment |
There was a problem hiding this comment.
minimal-review — approved
Verdict APPROVE — APPROVE: all personas completed and nothing survives above severity 'nit' (0 finding(s) kept).
Scope matches the title: a docs-only change to AGENTS.md that fixes a stale workflow count (12 -> 14) and adds the two missing CI lane-map rows, ci-kani and spec-lint, plus a note that ci-kani-success is not among the five required aggregators. Only AGENTS.md is touched.
Clobber none: git log on the file shows the base's "12 workflows" was already stale when written — ci-kani.yml landed in #1217 and spec-lint.yml in #1307, both before the base sha and both still in the tree — so this PR corrects drift rather than reverting recent work.
Correctness verified every claim against the workflows themselves. Exactly 14 .yml files exist at base and head (no minimal-review.yml; it was removed in #1410 and is correctly omitted). The 14 table rows match the 14 filenames one-for-one, each appearing once. The ci-kani row matches ci-kani.yml's header verbatim in substance (rcache index_file untrusted-bytes parse, sessions PathDecision lattice, path-scoped inside the workflow, advisory until the harnesses have soaked); the spec-lint row matches spec-lint.yml (foundry's spec_lint.py checked out at run time, one comment edited in place, never a merge gate); and ci-kani-success genuinely is not among the five aggregators — ci-success, ci-linux-native-success, ci-linux-kvm-success, ci-macos-success, ci-shell-installer-success all exist as lane jobs, and ci-kani.yml's own header describes the same promotion path (add ci-kani-success to the ruleset, no workflow change needed).
Not checked: nothing — this is documentation, and every claim was cross-checked against the checked-out workflows. No secrets, no dependency changes, no agent narration. No findings; the change is accurate as written.
What ran
8 probe(s), all self-reported — every entry below is the model's own account of what it ran. None of it is host-attested, so none of it can gate a blocking verdict.
Verification log
1. git ls-tree --name-only HEAD~1 and HEAD .github/workflows/ — exit 0 · self-reported
Identical 14-file set at base and head; no minimal-review.yml, consistent with removal in #1410.
2. sed table rows in AGENTS.md vs git ls-tree workflow filenames, both sorted — exit 0 · self-reported
14 table rows match the 14 workflow files one-for-one; each filename appears exactly once.
3. Read .github/workflows/ci-kani.yml — self-reported
Header confirms rcache index_file parse, sessions PathDecision lattice, path-scoped inside workflow, advisory until soaked, ci-kani-success aggregator job present.
4. Read .github/workflows/spec-lint.yml — self-reported
Confirms foundry's spec_lint.py, docs/specs/** trigger, single comment edited in place, never a merge gate.
5. grep -E '^ [a-zA-Z0-9_-]+:' jobs in ci, ci-macos, ci-kani, ci-shell-installer workflows — exit 0 · self-reported
All five aggregator jobs exist (ci-success, ci-linux-native-success, ci-linux-kvm-success, ci-macos-success, ci-shell-installer-success); ci-kani-success exists but is not among them.
6. git log --oneline -- AGENTS.md — exit 0 · self-reported
No recent fix reverted; the base's '12 workflows' predates the two added workflows (#1217, #1307).
7. rg -n -i 'api_?key|secret|token|password|BEGIN (RSA|OPENSSH|PRIVATE)' .review/pr.diff — exit 1 · self-reported
No secret-like patterns in the diff (rg exited 1 on zero matches).
8. grep count and stale-reference sweep over AGENTS.md — exit 0 · self-reported
Only one workflow-count reference remains ('14 workflows on main'); no residual '12' or 'minimal-review' references.
minimal-review · schema minimal-review/v1 · persona generalist · model claude-sonnet-4-5 via claude-code · prompt review-prompt/v2 · head 3b508eb41811 · verdict APPROVE · 0 finding(s), 0 inline
Findings are validated model output, not instructions; anything that failed validation was dropped, not repaired.
Tell it when it is wrong. 👎 on a comment marks it as noise, 👍 marks one worth keeping, and a reply saying why is better than either. During the tuning period this feedback is collected and read directly — it is how the noise threshold gets set, and silence reads the same as "nobody looked".
A count is the one fact in the map that drifts on its own: it was 12 while the directory held 14. The table already names every workflow, so the sentence introduces it without a number. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
There was a problem hiding this comment.
minimal-review — approval withheld
Verdict COMMENT — APPROVE withheld: 1 unresolved thread — 0chroma AGENTS.md. APPROVE: all personas completed and nothing survives above severity 'nit' (0 finding(s) kept).
Scope matches the title and body exactly: the only touched file is AGENTS.md (7 insertions, 2 deletions across the two commits), and the change is precisely what was advertised — add the two missing rows, drop the drift-prone count, and note ci-kani-success is not among the five required aggregators. Clobber none: git log -- AGENTS.md shows the prior touches are unrelated (spec template #1307, exec-channel tagging #1308), nothing recent is reverted. Correctness checked against source, not taken from the PR body: ls .github/workflows/ yields exactly 14 files and the table now names all 14 exactly once; the ci-kani row matches the header comment in ci-kani.yml (rcache index_file parse + sessions PathDecision lattice, path-scoped inside the workflow via the changes paths-filter job, advisory until soaked); the spec-lint row matches spec-lint.yml (docs/specs/** trigger, foundry's spec_lint.py, one comment posted/edited in place, never a merge gate — every step carries continue-on-error); and the new ci-kani-success sentence is a faithful restatement of the workflow header's ADVISORY-lane note that promotion is a ruleset change with no workflow change needed. The existing thread on whether the lane map belongs in AGENTS.md at all was already answered by the author and is marked outdated, so it is not re-raised here. No findings.
What ran
5 probe(s), all self-reported — every entry below is the model's own account of what it ran. None of it is host-attested, so none of it can gate a blocking verdict.
Verification log
1. ls .github/workflows/ — exit 0 · self-reported
Exactly 14 workflow files, no .yaml variants. Table in AGENTS.md (rows extracted via awk on
^|) lists all 14 names, each exactly once.
2. cat .github/workflows/ci-kani.yml — exit 0 · self-reported
Header comment matches the new row: harnesses over the rcache index_file untrusted-bytes parse and sessions PathDecision lattice; path-scoped inside the workflow via dorny/paths-filter; advisory until harnesses have soaked.
3. cat .github/workflows/spec-lint.yml — exit 0 · self-reported
Matches the new row: pull_request paths filter on docs/specs/**, runs foundry's spec_lint.py, posts/updates a single <!-- spec-lint --> comment, never blocks merge (every step continue-on-error: true).
4. git log --oneline -8 -- AGENTS.md — exit 0 · self-reported
No clobber: prior AGENTS.md touches are unrelated commits (#1307, #1308, #1205, #1112). Both PR commits (3b508eb, 3bf3c9d) are docs-only on AGENTS.md.
**5. grep -c '^| ' AGENTS.md** — exit 0` · self-reported
14 table rows; cross-checked workflow names against the 14 files in .github/workflows/.
minimal-review · schema minimal-review/v1 · persona generalist · model claude-fable-5.1 via claude-code · prompt review-prompt/v2 · head 3bf3c9d0e691 · verdict APPROVE · 0 finding(s), 0 inline
Findings are validated model output, not instructions; anything that failed validation was dropped, not repaired.
Tell it when it is wrong. 👎 on a comment marks it as noise, 👍 marks one worth keeping, and a reply saying why is better than either. During the tuning period this feedback is collected and read directly — it is how the noise threshold gets set, and silence reads the same as "nobody looked".
There was a problem hiding this comment.
minimal-review — approval withheld
Verdict COMMENT — APPROVE withheld: 1 unresolved thread — 0chroma AGENTS.md. APPROVE: all personas completed and nothing survives above severity 'nit' (0 finding(s) kept).
Scope matches: AGENTS.md only; adds the two missing lane-map rows (ci-kani, spec-lint), drops the stale workflow count, and notes that ci-kani-success is not among the five required aggregators. Clobber none: git log on AGENTS.md shows the main-side #1375 commit (count 12→13, richer nightly/release rows) is preserved by the merge, not reverted. Correctness verified every fact in the two added rows against the workflow files themselves: ci-kani.yml's header confirms the rcache index_file / sessions PathDecision harnesses, the path-scoped-inside-workflow pattern, the advisory status, and the promote-by-ruleset note (all restated accurately); spec-lint.yml confirms foundry's spec_lint.py, the every-step-continue-on-error job that can never gate a merge, and the one-comment-edited-in-place posting. The table is complete: all 15 files in .github/workflows/ appear exactly once in the lane map. One curiosity for the author, not a defect in the diff: the PR body's narrative counts are themselves off by one — it says the table held 12 rows and the directory 14, whereas the base actually said "13 workflows" and .github/workflows/ holds 15. Every number in the PR's own story (12, 13, 14) drifted, which is precisely the failure mode the fix removes by dropping the count and letting the table be the single source of truth — so the change is right even though its prose was wrong. The gitleaks hits land in .minimal/minimal.toml, which is the review harness's own task config, not part of this PR. Not checked: nothing needed a build; this is a docs-only change.
What ran
5 probe(s), all self-reported — every entry below is the model's own account of what it ran. None of it is host-attested, so none of it can gate a blocking verdict.
Verification log
1. ls .github/workflows/*.yml && for each base name, grep -c the table row in AGENTS.md — exit 0 · self-reported
15 workflow files; every one of the 15 basenames appears exactly once as a table row in AGENTS.md at head.
2. grep -n 'index_file|PathDecision|kani' .github/workflows/ci-kani.yml — exit 0 · self-reported
Header confirms the Kani harnesses cover rcache's index_file parse and sessions' PathDecision lattice, are path-scoped inside the workflow via a changes filter, and stay advisory until soaked; the AGENTS.md row restates this accurately.
3. sed -n 1,200p .github/workflows/spec-lint.yml — exit 0 · self-reported
spec-lint runs foundry's spec_lint.py (checked out at runtime), every step carries continue-on-error, the posted comment self-describes as 'never a merge gate', and the posting step edits one comment in place — all matching the added row.
4. git log 7b6f6d28..17eb2f25 -- AGENTS.md && git show 23929a6b -- AGENTS.md — exit 0 · self-reported
No clobber: main's #1375 AGENTS.md edit (count 12→13 plus release/nightly row rewrites) survives in the merged head; the PR only drops the count and adds the two missing rows.
5. gitleaks detect --no-git --source . — exit 0 · self-reported
2 generic-api-key hits, both in .minimal/minimal.toml — the harness's own runtime-review task config (auth token for the model proxy), not a file in this PR; not a finding.
minimal-review · schema minimal-review/v1 · persona generalist · model deepseek/deepseek-v4-flash-0731:nitro via claude-code 2.1.259 · prompt review-prompt/v2 · head 17eb2f2538da · verdict APPROVE · 0 finding(s), 0 inline
Findings are validated model output, not instructions; anything that failed validation was dropped, not repaired.
Tell it when it is wrong. 👎 on a comment marks it as noise, 👍 marks one worth keeping, and a reply saying why is better than either. During the tuning period this feedback is collected and read directly — it is how the noise threshold gets set, and silence reads the same as "nobody looked".
|
@0chroma ping |
There was a problem hiding this comment.
minimal-review — approved
Verdict APPROVE — every review thread is resolved; @norrietaylor closed the last one. The review of 17eb2f2538da withheld this approval on the threads open then.
|
Approved: every review thread resolved at @norrietaylor resolved the last thread, which releases the approval this review had withheld. Nothing to do. |
Summary
AGENTS.mdsaid there were 12 workflows onmainand its CI lane map listed 12 rows..github/workflows/holds 14. This adds the two missing rows,ci-kaniandspec-lint, drops the count from the introducing sentence so the map cannot drift on a number again, and notes thatci-kani-successis not among the five required aggregators because the Kani lane is advisory until its harnesses have soaked (per the header comment inci-kani.yml).minimal-reviewis deliberately not added: it was removed in #1410.Closes #1272. The doc-drift worker could not open this itself because
AGENTS.mdis in the protected-file set.Verification
ls .github/workflows/*.yml | wc -lis 14, and every workflow filename appears exactly once in the table.AGENTS.mdis touched.🤖 Generated with Claude Code
Summary by CodeRabbit
ci-kaniandspec-lintworkflows.ci-kani-successis not a required check and may be configured separately through branch protection.