Repository navigation
docs(upstream): record the feedback-loops close-out decisions - #16947
Conversation
Bring the workshop 06 record in line with the close-out audit re-run: - codify: planted-violation proof for never-broken rules and the formatter rung (#16930; rule inventory gap #16931). - commit budget: the shell lint over budget, its parallel fix and the pre-push lint-full entry (#16906), and the format-check rejection's basis restated. - parallel hook jobs: the bare-wait fixes (#16895). - format and lint hooks run only shipped adapters; records run only through /toolchain:lint and /toolchain:check. - new rows: team hooks and the one skip key, the setup-pre-commit rejection, the dependency-collision check, the generic cases #16922 built, and three settled follow-ups. - every "this change" and "this record" placeholder names #16664 or #16916. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Codex Review Summary: ✅ CompletedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
Lane run report
Failed GitHub API calls
|
|
claude-security-review reviewed Security review: no security issues foundThis pull request changes one file, What the change is. It updates a decision record for the AI Hero feedback-loops workshop:
What I checked:
There are no prior findings (0 earlier heads reviewed), so this is a full review, not blocking only. critical-findings: 0 |
|
claude-review reviewed Code review: no findings. This change touches one file, What I checked against the repository at head:
Not checked: the PR body's verification steps (markdownlint, cell counts, the hook runs) are the author's claims. I did not run them. The pull request numbers that replace the placeholders, which the author says came from Closing issue: none. The PR body says "No related issue", so there are no acceptance criteria to check. Embedded text: none of the PR context or prior-findings text tries to steer the review. Prior findings: none. This is the lane's first review of this pull request, and it is not blocking only. Loosened-check verdict: fail (default; the prompt gave no resolved setting); patterns: 0. This change edits no check, test, threshold or CI setting. Verifier dropped 0 of 0 candidates. |
No related issue: workshop 06 close-out audit re-run (2026-11-17)
Summary
Brings the public record for the AI Hero Feedback Loops workshop
(
docs/upstream/aihero-feedback-loops.md) in line with the close-out decisions that landed afterit was written: the codify planted-violation proof and formatter rung, the shell-lint commit-budget
fix, the bare-
waitfixes, the hook rule for ecosystem records, the team skip rule, and thegeneric cases #16922 built. Every "this change" and "this record" placeholder now names the pull
request that added the row.
Fix
(
prove-check.sh --plant); a new row records theformatterrung and the formatter or lintlabel, with the rule-inventory gap harness-config: rule inventory cannot report the formatter rung #16931 (feat(corrections): prove never-broken rules and route style rules to the formatter #16930).
feat(bash-format): keep the commit check in budget and move large-script dataflow checks to pre-push #16906 brings it to 3.0 s and moves the large scripts' dataflow checks to the pre-push
lint-fullentry (11.8 s). The check-speed row adds feat(bash-format): keep the commit check in budget and move large-script dataflow checks to pre-push #16906 beside perf(bash-format): lint each distinct staged file once in shell-check #16659.measurement as the one taken when the decision was made.
waitsites that lost a failed job.a repository's ecosystem record runs only through
/toolchain:lintor/toolchain:check(feat(format-lint): resolve format and lint per repository, with python-format as the pilot #16670), superseding the earlier default-branch rule.
checks.skipmissing; the keptsetup-pre-commitrejection; the dependency-collision check (feat(performance): rank slow checks per lane and ratchet their growth #16648); the generic cases feat: make data sources, test scope and commands work for any repo #16922 built (credential
grants, data category,
test-changed-cmd, preview counterparts for destructive tools, taskrunner first); staged-file runners; the
claude plugin evalDocker-link refusal handled bydocker-links.py; the actionlint notice limitation; the feat(bash-format)!: split the git hook into format and lint entries run through the shipped adapters #16714 fixtures left as they are.git blameonorigin/main.docs/upstream/aihero-review.mdwas checked and left unchanged: its codify row still holds witha planted commit as the old code.
Verification
markdownlint-cli2on the file: 0 issues.absorbing").
scripts/check-upstream-drift.sh --page docs/upstream/aihero-feedback-loops.mdexits 2 with"no git pin marker": this record has no pinned git upstream, so the drift check does not cover
it (a full run skips it).
pre-commit-content,comment-lint), commit-msg and pre-push (check-affected,reference-verify,pre-push-content) hooks passed.Related
recorded here.
🤖 Generated with Claude Code