mirzaaghazadeh/standbye/reviewing-code
Claude Code and agent skill by mirzaaghazadeh/standbye
- community
Checking that the SKILL.md file is still in the repository.
Review a teammate's change so the review is worth the time it costs — read the diff, run it, look for the bug that matters, and leave evidence rather than opinions. Use when reviewing a pull request, a branch or a patch.
Install
npx vskill@latest install mirzaaghazadeh/standbye/reviewing-codeSKILL.md
The same SKILL.md was published earlier as mirzaaghazadeh/standbye/reviewing-code.
Reviewing code
A review that says "looks good" is a review that did not happen. Your job is to find the thing that would have broken in production, and to say it in a way the author can act on in five minutes.
Read the change in the right order
- The description. What was this supposed to do?
- The tests. If the behaviour changed and the tests did not, that is your first comment.
- The diff itself, file by file, whole hunks — not just the changed lines. The line above the
change is often the one that breaks.
git logon the files touched, when something looks odd. Code that looks wrong is sometimes a
fix for a bug you have not heard about.
Then actually run it
Check out the branch, run the suite, and run the thing the change claims to do. A review with no command in it is a guess. Note the command you ran and what it printed — that is the evidence your comment rests on.
What to look for, in priority order
- Correctness. What input makes this wrong? Off-by-one, empty list, null, the error path, the
second call, two of them at once.
- Data and money. Anything that deletes, overwrites, force-pushes, sends mail, or spends.
These deserve a second read even when they look obvious.
- Security. Untrusted input reaching a shell, a query, a path, or a template. Secrets in the
diff or in a log line.
- The contract. Does this break a caller, a stored format, or an API someone depends on?
- Tests. Does the new test actually fail without the change? A test that passes either way is
decoration.
- Fit. Does it read like the rest of the repo?
Style you merely disagree with is not a finding. Say it once, mark it as optional, and let it go.
Writing the comment
- Be specific: file, line, the input that breaks it, and what you expect instead.
- Separate "this is broken" from "I would have done it differently". Authors ignore reviewers who
do not.
- Suggest the fix when you know it. It is faster for everyone than a Socratic question.
- If the change is good, say what you checked, so the author knows the review was real.
Post the review in the channel the team uses for it, and tell the author directly if something blocks the merge.
Your limits
Do not approve your own work. Do not merge on the author's behalf unless the owner has said you may. If the change is risky and you are unsure, that is exactly what ask_user is for — describe the risk in one line and recommend a call.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingpackages/supervisor/skills/reviewing-code/SKILL.md
Popularity
- Repository stars
- 6
- Repository forks
- 1
- 7-day trend score
- 15
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/mirzaaghazadeh/standbye/reviewing-code)