Repository navigation
docs(contributing): clarify verified CI requirement - #12763
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (3)
WalkthroughThe documentation updates define GitHub-verified signatures on every fork pull request commit as a prerequisite for automatic trusted-CI approval. The PR monitor guidance adds verification commands, summary text, and remediation instructions for unverified commits. ChangesFork CI verification
Estimated code review effort: 2 (Simple) | ~10 minutes 🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 inconclusive)
✅ Passed checks (4 passed)
Comment |
|
🔄 Datadog auto-retried 1 job - 1 passed on retry 🔗 Commit SHA: 673bd09 | Docs | Datadog PR Page | Give us feedback! |
c147828 to
0281aae
Compare
|
/ok to test 0281aae |
|
/ok to test c65121f |
harryskim
left a comment
There was a problem hiding this comment.
Approving — the added prose is accurate against .github/workflows/trigger-ci-approval-flow.yml, and the pages/reference/api/python/* churn is expected generator output.
Non-blocking follow-ups, happy for these to land separately:
- The English side got this in one place (
dco-and-licensing.md) while the zh-CN overview got it in two. The English counterparts of both zh-CN spots are untouched:getting-started.mdx(Configure DCO sign-offstep) andcontribution-flow.mdx(Complete automated review and CIstep).getting-started.mdxis the setup page contributors are routed to, so it's the highest-value place for the note. /ok to testappears nowhere else indocs/fern/pages, so the new section references an undefined term for English readers. The zh-CN overview defines it, which is why it reads fine there.- Two things the workflow does that the docs don't say: signing alone isn't sufficient (a contributor not in
APPROVED_EXTERNAL_CONTRIBUTORSnever gets automatic approval regardless of signatures), and a maintainer can still post/ok to test <sha>manually — the bot's own comment tells contributors they don't need to rewrite commits for that path. Without #3, expect "I signed everything, why no CI?" reports and unnecessary branch-wide re-signing.
Minor: the new section splits the page's Create -> Verify -> Repair spine ("Repair the Latest Commit" now reads as if it repairs the signing problem); Verified is bold in CONTRIBUTING.md but backticked elsewhere; and the Devin comment about the DCO page not being updated is stale.
c65121f to
1b13726
Compare
Signed-off-by: Kristina Pathak <kristina.m.pathak@gmail.com>
Signed-off-by: Kristina Pathak <kristina.m.pathak@gmail.com>
Signed-off-by: Kristina Pathak <kristina.m.pathak@gmail.com>
1b13726 to
673bd09
Compare
|
/ok to test 19cd52f |
dagil-nvidia
left a comment
There was a problem hiding this comment.
Approving: all gates green at head 19cd52f (skills validation proven with a known-positive, CODEOWNERS drift 0, strict coverage, 93 matcher tests, docs linter proven with an injected-H1 probe), and the governance wording was fact-checked clause by clause against .github/workflows/trigger-ci-approval-flow.yml, the ground truth: the fork-only condition, all three approval paths, the every-commit-Verified requirement, the signature-check-after-approval ordering ("signing commits does not create that path" is exactly right), the unsigned-commit comment, and the manual /ok to test <sha> fallback all match. Every added sentence keeps the external-contribution gate closed, and the scope constraint from issue #12761 is honored on all five English surfaces plus the zh-CN page. The earlier review's follow-ups are all addressed at this head.
Three non-blocking notes:
.agents/skills/pr-monitor/SKILL.md:57- the documented verification check makes N+1 API calls;.commit.verificationis already in the commits-list response. Verified live on this PR that the one-call form returns identical output:gh api --paginate "repos/ai-dynamo/dynamo/pulls/$PR_NUMBER/commits?per_page=100" --jq '.[] | "\(.sha[0:7]) verified=\(.commit.verification.verified) reason=\(.commit.verification.reason)"'.AGENTS.md:148- "the automatic approval flow posts that command" refers back to a definition that says "with the short SHA", but the workflow posts the full 40-character head SHA. Behavior unaffected; the referent reads wrong. Suggest "...posts the same command with the full head SHA only after every PR commit is GitHub-verified.".agents/skills/pr-monitor/SKILL.md:50- reason 3 packs flow, requirement, diagnosis, fix, and fallback into one five-sentence bullet while its siblings are one-liners; keep the reason line short and move the detail into the preceding paragraph.
Pre-existing, untouched lines: the internal-looking nvevents.nvidia.com links the linter warns about, and the structural EN/zh divergence of the legacy zh-CN overview page (updating it in place was the only option).
Automated review pass; severity escalation stays with the human reviewers.
|
/ok to test 3a82369 |
|
Is this targeted for v1.5.0? Code freeze is today at 3:00 PM PT - to make the release it needs to merge to main by then. |
|
/ok to test 3a82369 |
Signed-off-by: Kristina Pathak <kristina.m.pathak@gmail.com> Co-authored-by: Dan Gil <dagil@nvidia.com>
Summary
pr-monitortroubleshooting.Validation
git diff --checkuv run --no-project --with PyYAML python3 scripts/validate_skills.pySummary by CodeRabbit