- Home
- Skills
- Tsaol
- Awesome Claude
- Git Review
git-review_skill
- Python
44
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
3 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill tsaol/awesome-claude --skill git-review- SKILL.md11.6 KB
Overview
This skill reviews code changes against stated requirements and the original design, validates implementation correctness, and runs tests before commit or PR. It produces a structured report with a verdict, mapped requirements coverage, design alignment notes, code-quality findings, and test results. Use it as a final sanity and quality gate to catch regressions, scope drift, and security/logic issues.
How this skill works
The skill gathers context from uncommitted diffs, a provided issue/PR URL, commit messages, and project design documents or inferred architecture. It then maps code changes to requirements, checks design alignment and naming/pattern consistency, runs the project test suite when requested, and classifies findings by severity. Finally it generates a concise review report with actionable fixes and recommended next steps.
When to use it
- Before committing changes or opening a pull request
- After implementing a feature to verify requirements were met
- When validating a patch against an issue or PR description
- To check for unintended scope creep or unrelated edits
- Before running /git-commit as a final quality gate
Best practices
- Provide an issue or PR URL or a short requirements description for accurate mapping
- Run the review iteratively: fix high-severity items and re-run the tool
- Include design docs (DESIGN.md, ARCHITECTURE.md) in repo or rely on consistent project structure
- Prefer small, focused commits to make mapping requirements to changes easier
- Add tests for new behavior and re-run review with --test-only
Example use cases
- Review uncommitted changes for a completed feature to ensure acceptance criteria are implemented
- Validate a branch against a GitHub issue URL to confirm alignment before creating a PR
- Run design-only checks when tests are slow or unavailable to focus on architecture and API compatibility
- Run test-only checks after adding tests or fixing a failing CI job
- Detect scope creep by flagging changes that don't map to stated requirements
FAQ
It works with uncommitted diffs by default and accepts a GitHub issue or PR URL or a short user-provided requirements description for mapping.
Can I skip tests or only run tests?
Yes. Use --design-only to skip tests and focus on requirements/design, or --test-only to run only the test phase and report results.