PR guardian

PR Guardian is an AI-powered code review system designed to make software reviews faster, more comprehensive, and easier to understand. Developers can provide a GitHub Pull Request or local project files, and PR Guardian automatically analyzes the code using multiple specialized AI agents. The system uses separate agents for Code Quality, Security, Testing, and Performance. These agents run in parallel, allowing different aspects of the code to be reviewed simultaneously. The Code Agent identifies logic errors, unhandled exceptions, resource issues, and code-quality problems. The Security Agent searches for vulnerabilities such as injection risks, exposed secrets, insecure authentication, XSS, and sensitive-data exposure. The Testing Agent identifies missing test coverage, edge cases, and potentially failing tests. The Performance Agent looks for inefficient algorithms, blocking operations, unnecessary database queries, memory issues, and other bottlenecks. The findings are consolidated into a structured report containing the affected file, severity, explanation, problematic code, confidence level, and suggested solution. PR Guardian also calculates an overall risk level across security, code quality, performance, and testing. A Fix Agent can generate complete corrected files based on detected issues. When required, fixes can be validated using automated regression testing before being presented to the developer. For GitHub projects, developers can optionally commit approved fixes directly back to the relevant branch. The system is designed to reduce manual review effort while giving developers actionable explanations rather than simply reporting errors. Its modular multi-agent architecture allows additional review agents and analysis capabilities to be added in the future.