andreaswasita/copilot-agents-dojo/receiving-code-review

Receiving Code Review Skill

Processes every review comment — human, agent, or CI — until the change is approved. Each comment is categorized, addressed, verified, and answered. Does NOT silently ignore comments and does NOT amend the original commit during active review.

When to Use

• Review comments received on a pull request.
• User provides corrections or improvement suggestions.
• The requesting-code-review skill flagged issues.
• CI/CD checks fail with actionable feedback.
• A sub-agent's review output needs to be addressed.

Prerequisites

• The list of comments (PR thread, CI log, or chat).
• The view, edit, grep, and powershell Copilot tools.
• git for new commits and gh for PR threads (when relevant).
• A green baseline test run.

How to Run

1. Read every comment. Categorize by action.
2. For each comment: fix, verify, respond, mark resolved.
3. Re-run the full test suite after all fixes.
4. Re-self-review with the `requesting-code-review` skill.
5. Update `tasks/todo.md` and (if pattern recurring) `tasks/lessons.md`.
6. Request re-review with a short summary.

Quick Reference

Procedure

Step 1: Acknowledge and Categorize

Read every comment. Categorize each using the table above. Track them in tasks/todo.md as a Review Feedback section:

### Review Feedback (Round 1)
-[ ] Comment #1: missing error handling — agree, will fix
-[ ] Comment #2: rename for clarity — agree, will fix
-[ ] Comment #3: add pagination — agree, follow-up (out of scope)

Step 2: Address Each Comment

For every comment:

1. Fix the issue with edit.
2. Run the relevant tests via the powershell tool.
3. Respond on the thread with what changed and why (cite commit SHA).
4. Tick the checkbox in tasks/todo.md.

### Comment #1: missing error handling — @reviewer
**Action:** Added try/catch with 502 response on upstream failure (commit `abc1234`).
**Verification:** New test covers the failure path.

Step 3: Re-Verify After All Fixes

Run the full test suite — not just the tests for the changed lines. Re-run the dojo gate:

bash scripts/verify.sh --check

Re-self-review with the requesting-code-review skill. Every fix is a potential new bug.

Step 4: Record Lessons If The Pattern Repeats

If review keeps catching the same shape of issue, append to tasks/lessons.md:

- date: 2026-05-19
  error_type: missing-error-handling
  trigger: "Reviewer caught missing try/catch on external API call"
  root_cause: "Did not consider failure modes for external dependencies"
  fix: "Wrapped call in retry + circuit breaker"
  rule: "Every external API call has an explicit failure path with a test"
  occurrences: 2
  status: active

Step 5: Request Re-Review

**Re-review ready**
-Addressed 2/3 comments directly
-1 logged as follow-up (out of scope)
-All tests pass (49 total, 2 new)
-Self-review clean

Push fixes as new commits while review is active. Squash fixup commits just before merge — not during.

Pitfalls

• DO NOT ignore a comment. Every comment gets a response, even "won't fix, here's why".
• DO NOT be defensive. "It works in my tests" is not a response to "this could be better".
• DO NOT force-push during active review. Reviewers lose context.
• DO NOT fix one comment in isolation and skip retesting. Re-verify after every round.
• DO NOT treat repeat feedback as random. If the same theme keeps coming back, log a lesson.

Verification

• [ ] Every comment is checked off in tasks/todo.md.
• [ ] Every comment has an explicit thread response (action + commit reference).
• [ ] Full test suite re-run after all fixes; exit 0.
• [ ] requesting-code-review re-run after fixes; no new 🔴 findings.
• [ ] If a pattern repeated, tasks/lessons.md was updated.