agents-review
Review code changes using src.agents.scripts.review after edits. Use when you need multi-provider feedback on diffs, staged changes, or the last commit, or when focusing on a review theme like security.
When & Why to Use This Skill
The Agents Review skill automates the code analysis process by providing multi-provider feedback on various code states, including working trees, staged changes, and specific commits. It enhances code quality and security by allowing developers to apply specialized review themes, ensuring that every diff is scrutinized for potential bugs, vulnerabilities, and architectural consistency before finalization.
Use Cases
- Pre-commit Validation: Automatically review staged changes to identify logic errors or style violations before they are committed to the repository.
- Security Auditing: Focus the review process specifically on security vulnerabilities by applying a security-themed analysis to recent code edits.
- Continuous Integration Support: Integrate multi-provider feedback into the development workflow to ensure high-quality code standards across the team.
- Post-Commit Analysis: Quickly evaluate the impact and quality of the last commit (HEAD) to facilitate rapid iterations and debugging.
| name | agents-review |
|---|---|
| description | Review code changes using src.agents.scripts.review after edits. Use when you need multi-provider feedback on diffs, staged changes, or the last commit, or when focusing on a review theme like security. |
Agents Review
Overview
Use this skill to run multi-provider review after changes, especially before finalizing or committing.
Quick start
Review all changes in the working tree.
uv run python -m src.agents.scripts.review
Common scopes
Review only staged changes.
uv run python -m src.agents.scripts.review review.scope=staged
Review the last commit.
uv run python -m src.agents.scripts.review review.scope=head
Focus the review on a theme like security.
uv run python -m src.agents.scripts.review 'review.focus=security'