What problem does it solve?
Developers frequently miss critical code defects, security vulnerabilities, and quality regressions in their own work before committing, leading to costly post-commit fixes, security incidents, and broken production builds. Manual code review is time-consuming, inconsistent, and often fails to catch subtle issues that automated tools and fresh reviewer context can identify immediately.
Core Features & Use Cases
- Automated Static Security Scanning: Detects common vulnerabilities like hardcoded secrets, shell injection, SQL injection, and unsafe deserialization in added code lines before commit.
- Baseline-Aware Quality Gates: Only flags new test failures or lint errors introduced by changes, ignoring pre-existing issues in the codebase to avoid false positives.
- Independent Subagent Review + Auto-Fix: Dispatches a separate, context-isolated reviewer agent to catch logic errors and security issues the implementer missed, with an automated fix loop that resolves reported problems in up to two retry cycles.
- Use Case: A developer finishing a new payment processing feature can run this skill to automatically scan for security flaws, run tests to check for regressions, get an independent review, and fix any minor issues before committing, preventing vulnerable code from reaching the repository.
Quick Start
Use the requesting-code-review skill to verify your recent code changes and automatically fix any identified issues before committing them to your git repository.
