From harness-engineering
Analyze, respond to, and fix PR review comments, then save learnings to memory. Use when review comments arrive on a PR, when asked to address review feedback. Do NOT use for reviewing a PR yourself (use pr-reviewer-3a), pre-PR checks (use quality-gate).
npx claudepluginhub toru-oizumi/claude-harness-engineering --plugin harness-engineeringThis skill uses the workspace's default tool permissions.
Process review comments on a pull request end-to-end: analysis, response, fixes, and learning capture.
Generates design tokens/docs from CSS/Tailwind/styled-components codebases, audits visual consistency across 10 dimensions, detects AI slop in UI.
Records polished WebM UI demo videos of web apps using Playwright with cursor overlay, natural pacing, and three-phase scripting. Activates for demo, walkthrough, screen recording, or tutorial requests.
Delivers idiomatic Kotlin patterns for null safety, immutability, sealed classes, coroutines, Flows, extensions, DSL builders, and Gradle DSL. Use when writing, reviewing, refactoring, or designing Kotlin code.
Process review comments on a pull request end-to-end: analysis, response, fixes, and learning capture.
gh pr view <PR_NUMBER> --json number,title,headRefName
gh api repos/<OWNER>/<REPO>/pulls/<PR_NUMBER>/comments
gh api repos/<OWNER>/<REPO>/pulls/<PR_NUMBER>/reviews
For each comment, run proposal-evaluator to produce a principled verdict before acting.
Do not assume every comment requires a code change.
Quick triage first — classify to decide whether to evaluate or fast-track:
| Category | Description | Next step |
|---|---|---|
| Already resolved | Addressed in a later commit | Reply "Resolved in <commit>", skip evaluator |
| Factual fix | Clear bug or broken code with no ambiguity | Accept immediately, note rationale |
| Needs evaluation | Opinion, suggestion, stylistic, or disputed | Run proposal-evaluator |
For "Needs evaluation" comments, run the full 5-axis evaluation:
See
proposal-evaluatorskill for the scoring rubric and verdict template.
The evaluation produces one of:
Priority order: Already resolved → Factual fix → Evaluate others
For each Factual fix comment and each Accept or Counter verdict from Step 2:
git add .)fix: address review comment - <short description>
Reply to each comment on GitHub:
gh api repos/<OWNER>/<REPO>/pulls/<PR_NUMBER>/comments/<COMMENT_ID>/replies \
-f body="<reply text>"
Reply patterns:
<commit-sha>. [Brief explanation of what changed]"After handling all comments, abstract reusable patterns and save to memory:
What to save:
What NOT to save:
git push origin <branch>
| User Input | Action |
|---|---|
Handle PR comments | Execute all steps for current branch PR |
Handle PR comments #123 | Execute all steps for PR #123 |
Reply to comments only | Steps 2 and 4 only (no code changes) |
proposal-evaluator — 5-axis evaluation framework used in Step 2See gotchas.md in this directory for known pitfalls and recurring mistakes when using this skill.