Autonomous Multi-Agent Code Review Swarm
Architect a supervisor-worker AI agent network for automated PR review and linting.
Problem Statement
Engineering teams lose hours conducting repetitive code reviews. In this project, you will build an autonomous agent swarm where specialized LLMs (Security, Performance, and Architecture) inspect git diffs.
What You'll Build
A supervisor-worker AI agent network that receives GitHub pull_request webhooks, orchestrates parallel subagent reviews, and writes inline annotations directly onto code lines.
Technology Stack
Skills You'll Master
Requirements & Prerequisites
- Python 3.11+ and virtual environment setup.
- API key from OpenAI, Gemini, or local Ollama endpoint.
Guided Milestones (3)
Step-by-Step ArchitectureGit Diff Parser & Context Tokenizer
Parse unified diffs into clean structured file blocks with line number mapping.
- Diff parser module
- Token-budgeted prompt builder
LangGraph Multi-Agent Architecture
Construct supervisor graph that delegates chunks to Security and Performance agents.
GitHub Webhook & Bot Feedback Loop
Deploy FastAPI webhook handler that posts structured GitHub check runs.
Project Submission & Verification
Submit your public GitHub repository and live deployment link for automated test checks.
Add to ASCI Portfolio
Demonstrates practical production-grade AI agent engineering beyond simple chatbot wrappers.