code-reviewer

Automated code review identifies security issues in pull requests by applying static analysis, pattern detection, and quality gates. It evaluates source files for hardcoded credentials, SQL injection, and command injection, generating explicit PASS or FAIL outcomes based on severity.

CWE scanning works by mapping detected vulnerabilities to common weakness enumerations during a security audit. It analyzes source files for path traversal and command injection risks, generating severity-aware reports that flag specific security flaws before generating a final PASS or FAIL outcome.

You can review code using bilingual requests in both Chinese and English. The workflow processes review instructions in both languages while maintaining consistent security references, CWE mappings, and defect reporting across the generated outputs.

When quality gates fail during a code review workflow, the process stops and triggers a rollback-aware recovery. This prevents merging changes by halting the review cleanly, allowing reviewers to act on severity-aware defects without losing process control.

Preventing quality regressions requires applying structured quality gates and static analysis to pull requests. By enforcing severity-aware reporting and rollback handling, the workflow detects defects and security issues early, ensuring only validated changes pass the final review.