skills/screening-pr-review/SKILL.md
PR screening review before human review. Use when: スクリーニングレビュー, PRレビュー支援, レビュー前チェック, PR screening, pre-review.
npx skillsauth add thkt/dotclaude screening-pr-reviewInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
| Step | Action | Purpose | | ---- | --------------------------- | -------------------------- | | 1 | Gather PR context | Background and intent | | 2 | Summarize changes per file | Before/after understanding | | 3 | Check dependency impact | Regression risk | | 4 | Produce findings + comments | Reviewer-ready output |
# Metadata
gh pr view --json title,body,labels,files,url $PR
# Diff
gh pr diff $PR
# Existing comments
gh pr view --comments $PR
# Inline comments
gh api repos/{owner}/{repo}/pulls/{number}/comments \
--jq '.[] | {file: .path, user: .user.login, comment: .body}'
Never include author in gh output fields.
| Label | Meaning | Severity |
| -------- | ------------------------------- | -------- |
| [must] | Requires fix before merge | High |
| [want] | Should fix, not blocking | Medium |
| [imo] | Personal opinion, take or leave | Low |
| [ask] | Question needing clarification | - |
| [nits] | Minor style/formatting issue | Low |
| [info] | Context sharing, no action | - |
| Rule | Detail |
| --------------- | ------------------------------------------------------------------------------------ |
| Format | [label] <observed behavior or risk>. <suggestion>. (file:line) |
| Concise | 3 lines for [imo]/[nits]/[info]; up to 5 for [must]/[want] with evidence |
| Respectful | Acknowledge effort, avoid commands |
| Suggestive | "Consider..." not "This is wrong" |
| Author-targeted | Comments may be posted verbatim — calibrate detail for the PR author |
| Topic | File |
| ---------------- | ---------------------------------------------------- |
| Review Checklist | ${CLAUDE_SKILL_DIR}/references/review-checklist.md |
tools
Delegate implementation to codex (coder) via the herdr-agentchat plugin and drive a two-pane conversation to completion.
development
Extract recurring patterns from past closed PRs/issues and the research findings in workspace/research/, verify them against the latest code, and propose them to docs/wiki/ via PR.
development
Create Decision Records (DR) in MADR v4 format with auto-numbering.
development
Detect flaky tests by shaking them — repeated runs under varied order, parallelism, and seed — plus a static smell scan that flags latent flakiness in tests that currently pass. Classify each target as confirmed-flaky, latent-flaky, or stable and fix the root cause without weakening the test. Do NOT use to fix a confirmed single bug (use /fix) or for static-only code review (use /audit).