/full-review

Run a complete review pipeline: agent-review first, then check-pr. The agent-review pass naturally fills the ~4 minute Copilot review delay, so check-pr starts with comments already waiting.

Arguments

• $ARGUMENTS - PR number (optional, defaults to current branch's PR)

Instructions

Phase 1: Agent Review

Run the /agent-review skill on the PR. This is a deep expert review that:

• Reads CLAUDE.md and the full PR diff
• Reviews against project-specific code quality, architecture, and testing criteria
• Posts a review comment on the PR
• Creates follow-up issues for deferred suggestions
• Reconciles any from-review issues resolved by this PR

Capture the results: verdict, findings counts, issues created/closed.

Phase 2: Check-PR

After agent-review completes, run the /check-pr skill on the same PR. By now, Copilot review has typically arrived (~4 min). This skill:

• Waits for Copilot review if still pending (Step 0 polling)
• Processes every review comment (Copilot + human + agent-review findings if inline)
• Fixes, dismisses, or defers each comment with inline replies
• Pushes all fixes and verifies every thread has a reply
• Resolves every conversation thread via GraphQL so branch protection's "conversations resolved" gate clears. Replies alone don't do this.
• Cross-references fixes against open from-review issues

Capture the results: comments processed, fixes committed, issues created/closed.

Phase 2.5: Verify CI (Optional)

If check-pr pushed any fix commits in Phase 2, CI needs to pass on the new HEAD before merge. Concurrency groups commonly cancel the in-progress run when fixes are pushed, leaving CI stale.

1. Check if any commits were pushed in Phase 2 (check-pr fixes)
2. If yes, run /fix-ci on the same PR
3. Common outcome: retriggering a cancelled run after concurrency cancellation
4. If no commits were pushed, skip this phase (CI is still valid from before)

Capture the results: CI status, any action taken (retrigger/fix/escalate).

Phase 3: Combined Summary

Output a single combined summary table covering both phases. This is the PRIMARY output.

| PR | Review | Check-PR | CI | Changes | Issues |
|----|--------|----------|----|---------|--------|
| #XX | Verdict (N critical, M suggestions) | P comments → Q fixed | PASS (after retrigger) | brief change 1, change 2 | Created: #A, #B. Closed: #C, #D |

Column guide:

• Review: Verdict + finding counts from agent-review
• Check-PR: N comments → M fixed (add , X false pos / , Y deferred if any)
• CI: Status from Phase 2.5. PASS / PASS (after retrigger) / PASS (after fix) / ESCALATED / — (if Phase 2.5 was skipped because no commits were pushed)
• Changes: Comma-separated brief descriptions of what changed (2-5 words each, from check-pr fixes)
• Issues: Combined from both phases. Created: #X for new follow-ups. Closed: #Y for resolved issues. Deduplicate (agent-review may create issues that check-pr then closes).

Then below the table:

• Full commit hashes for each fix
• Reasons for any false positives
• URLs for all created/closed issues
• PR ready for re-review: Yes/No

Execution Notes

• Sequential, not parallel. Agent-review MUST complete before check-pr starts. This is by design — the delay lets Copilot review arrive.
• Same branch. Both skills operate on the same PR branch. Check-pr may commit fixes on top of the reviewed code.
• Deduplication. If agent-review creates a follow-up issue and check-pr's fixes resolve it, close the issue in Phase 2 with a PR cross-reference.
• Threads resolved before declaring done. Check-pr's step 6b runs the GraphQL resolveReviewThread mutation for every thread. Without it, branch protection blocks merge silently — the user has to click "Resolve conversation" once per thread. If you skip this, full-review is not done; you've handed the user manual cleanup.
• Attribution. Follow Zero Attribution Policy throughout — no AI mentions in commits, replies, or issues.