marsonshine/ai-skills/review-code-change
Claude Code and agent skill by marsonshine/ai-skills
- community
Checking that the SKILL.md file is still in the repository.
Review an existing PR, commit, patch, diff, or changed files for concrete defects and regressions. Not for implementation, whole-repository audits, or style-only cleanup.
Install
npx vskill@latest install marsonshine/ai-skills/review-code-changeSKILL.md
Review Code Changes
Review the actual change against its repository, not against a generic style guide. A short review with one substantiated defect is better than a long list of preferences.
Workflow
- Establish the exact scope: live base, exact head, dirty files, and any retargeting or rebasing since the request began.
- Read applicable repository instructions and the contracts or testing rules relevant to the changed path. Load architecture records only when they bear on a suspected defect or constraint.
- Inspect the changed lines plus enough callers, callees, schemas, configuration, and tests to understand the behavior.
- Trace important paths end to end, including failures, cancellation, cleanup, concurrency, persistence, and alternate entry points.
- Verify claims with focused tests, static checks, logs, or reproducible reasoning from the code. Do not assume a green test suite proves semantic correctness.
- Report only actionable findings. Separate defects from optional improvements and disclose unverified areas.
Read references/review-dimensions.md when the change touches concurrency, durable state, permissions, public APIs, generated output, or runtime integration.
Specialist Routing
- For C#, ASP.NET Core, EF Core, or .NET-specific review, load
../csharp-dotnet-code-checklist/SKILL.mdin addition to this skill. - Prefer an available language, framework, security, database, or platform specialist when it provides narrower rules.
- Keep this skill responsible for cross-cutting review quality and reporting.
Reporting
Lead with findings ordered by impact. For each finding, state the location, defect, consequence, evidence, and smallest credible correction. If no findings survive verification, say so and list residual risks or checks you could not perform.
Do not report style nits already enforced by a passing formatter or linter unless they conceal a behavioral problem.
Report a verification gap only when a concrete behavior risk or required gate lacks credible evidence; the absence of new tests alone is not a defect.
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 repositoryCheckingskills/review-code-change/SKILL.md
Popularity
- 7-day trend score
- 0
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/marsonshine/ai-skills/review-code-change)