RepoForge AI: Agentic Test Generation for Public Python Repositories

A multi-agent system that scans public Git repositories and generates validated pytest suites using RAG over ChromaDB-indexed code with Gemini-based generation.

Built a multi-agent system (an OrchestratorAgent plus specialised agents for ingestion, code analysis, API discovery, retrieval, test planning, and test writing) that scans public Git repositories and generates validated pytest suites.

Combined RAG over ChromaDB-indexed code chunks with Gemini-based generation, plus safety checks (secret redaction, unsafe-pattern blocking) to produce reviewable, downloadable test artifacts.