|
| 1 | +--- |
| 2 | +name: github-pr-review-comments |
| 3 | +description: Use when the user asks to review GitHub pull requests, identify concrete risks, and write review comments with `gh`. Focus on actionable findings, not summaries. |
| 4 | +--- |
| 5 | + |
| 6 | +# GitHub PR Review Comments |
| 7 | + |
| 8 | +## When To Use |
| 9 | + |
| 10 | +Use this skill when the user asks to: |
| 11 | + |
| 12 | +- review open GitHub PRs |
| 13 | +- check PRs not opened by the repo owner |
| 14 | +- identify bugs, regressions, or missing validation in a PR |
| 15 | +- write review comments back to GitHub with `gh` |
| 16 | + |
| 17 | +## Goal |
| 18 | + |
| 19 | +Produce high-signal PR review findings and, when requested, post concise review comments with `gh`. |
| 20 | + |
| 21 | +## Review Standard |
| 22 | + |
| 23 | +Default to a code review mindset: |
| 24 | + |
| 25 | +- prioritize correctness bugs |
| 26 | +- prioritize behavioral regressions |
| 27 | +- prioritize compatibility and rollout risk |
| 28 | +- prioritize missing validation where the blast radius is meaningful |
| 29 | +- avoid low-value style commentary unless it masks a real bug |
| 30 | + |
| 31 | +A review comment should only be posted if it says something concrete that could block, regress, or materially weaken the change. |
| 32 | + |
| 33 | +## Workflow |
| 34 | + |
| 35 | +1. Identify candidate PRs |
| 36 | +- Prefer `gh pr list --repo friuns2/codexUI --state open`. |
| 37 | +- If the user asks for PRs not by owner, filter out PRs whose author matches the repo owner. |
| 38 | + |
| 39 | +2. Gather PR context |
| 40 | +- Read PR title, author, body, changed files, and commits: |
| 41 | + - `gh pr view <number> --repo friuns2/codexUI --json ...` |
| 42 | +- Read the diff: |
| 43 | + - `gh pr diff <number> --repo friuns2/codexUI` |
| 44 | +- Narrow attention to risky files and changed contracts. |
| 45 | + |
| 46 | +3. Form findings |
| 47 | +- Each finding should answer: |
| 48 | + - what can break |
| 49 | + - why it can break |
| 50 | + - where in the patch it comes from |
| 51 | +- Prefer 0-3 real findings over a long weak list. |
| 52 | + |
| 53 | +4. Write comments when requested |
| 54 | +- Use `gh pr comment <number> --repo friuns2/codexUI --body ...` |
| 55 | +- Keep comments short, direct, and technical. |
| 56 | +- One comment per distinct issue is preferred when the issues are unrelated. |
| 57 | + |
| 58 | +## Comment Style |
| 59 | + |
| 60 | +Use a terse, high-conviction review style: |
| 61 | + |
| 62 | +- direct |
| 63 | +- unsentimental |
| 64 | +- technical |
| 65 | +- specific about the failure mode |
| 66 | + |
| 67 | +Good pattern: |
| 68 | + |
| 69 | +1. Name the concern in one sentence. |
| 70 | +2. Explain the concrete breakage path. |
| 71 | +3. State what needs to be clarified, tested, or fixed before merge. |
| 72 | + |
| 73 | +Avoid: |
| 74 | + |
| 75 | +- excessive hedging |
| 76 | +- praise padding |
| 77 | +- long summaries of the PR before the issue |
| 78 | +- vague “maybe consider” language when the risk is real |
| 79 | + |
| 80 | +## Repo-Specific Notes |
| 81 | + |
| 82 | +- Canonical GitHub repo: `friuns2/codexUI` |
| 83 | +- Prefer reviewing non-owner PRs first when the user asks broadly. |
| 84 | +- If a finding is platform-related, explicitly call out the platform that should be tested, for example WSL, Linux, macOS, or Windows. |
| 85 | +- If you post comments, preserve the user’s requested tone as long as it stays professional and non-abusive. |
| 86 | + |
| 87 | +## Output Expectations |
| 88 | + |
| 89 | +When reporting back to the user: |
| 90 | + |
| 91 | +- list findings first |
| 92 | +- include PR number |
| 93 | +- include the short reason |
| 94 | +- mention whether comments were posted |
| 95 | + |
| 96 | +If no real findings are found, say so directly. |
0 commit comments