# Axion

- **Event:** [AIE Code Agents Hackathon](https://cerebralvalley.ai/e/aie-code-agents-nyc)
- **When:** Nov 22 at 9:00 AM – Nov 23 at 5:00 PM (EST)
- **Where:** The Malin Flatiron, New York, NY
- **Team:** [Kiitan Ayandosu](https://cerebralvalley.ai/u/kiitan), [Clinton Imaro](https://cerebralvalley.ai/u/clinton), [Emmanuel Mokua](https://cerebralvalley.ai/u/imanmokua), [Oluwasegun Soji-John](https://cerebralvalley.ai/u/segunsj)
- **GitHub:** https://github.com/clintonimaroo/agl ,  https://github.com/clintonimaroo/agl-backend
- **Demo video:** https://youtu.be/2PxZoV0?si=o0gI6zIzL0hk
- **Gallery:** https://cerebralvalley.ai/e/aie-code-agents-nyc/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/aie-code-agents-nyc/hackathon/gallery/28

Axion is a constitution driven multi agent engineering system that lets AI work inside real codebases with structure and safety. Modern AI tools generate code but do not understand architecture or hidden rules. Axion fixes this with a Constitution for Codebases, a formal rule set that guides every agent decision.

Axion uses a team of expert agents that collaborate like real engineers. They review each other, flag risks, call in specialists, and let the user join the conversation at any time. A sandbox environment enforces rules, tracks state, and ensures only safe diffs reach the repo.

Axion solves Problem Statement 1 by giving models persistent context through the constitution and by assigning expert roles so agents reason with stable knowledge of the codebase. It solves Problem Statement 3 by enabling real multi agent and human collaboration. Axion turns natural language tasks into clean, architecture aware code changes.

---

Markdown version of https://cerebralvalley.ai/e/aie-code-agents-nyc/hackathon/gallery/28. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt