CodeReview-AI-Agent
About
CodeReview-AI-Agent was built as a Kaggle Agents Intensive 2025 capstone on Google's Agent Development Kit, structuring review as three cooperating agents: a Code Analyzer, a Security Checker, and a Quality Reviewer, coordinated by an orchestrator with shared session state. Agents run sequentially or in parallel depending on configuration, with parallel execution cutting review latency substantially on multi-language changes. The pipeline covers Python, JavaScript, TypeScript, Java, Go, and Rust, combines LLM judgment with Python AST analysis, and produces reports in HTML, Markdown, SARIF, and JSON — SARIF enabling direct IDE and GitHub Security integration. A GitHub integration posts review comments and inline annotations on pull requests, and a GitHub Actions workflow runs reviews from CI. The repository is a capstone portfolio project with 24 commits, not an actively maintained production tool.