# Kamel Elshikh

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Kamel Elshikh](https://cerebralvalley.ai/u/Kamel2026)
- **GitHub:** https://github.com/kamel1883-ops/aegiscode-ai
- **Demo video:** https://youtu.be/q87hH0ZmlSE
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/114

AegisCode AI is an autonomous cyber-vulnerability detection and self-healing security agent designed to identify, analyze, and patch software security flaws automatically.

Powered by Meta's Llama models, the system operates using a multi-agent architecture integrated with Docker container isolation and a 4-condition causal verification matrix. Instead of merely alerting developers to security risks, AegisCode AI parses AST context, detects vulnerabilities, generates precise context-aware patches, and verifies that the fixes resolve the issue without introducing regressions or breaking existing code functionality.

## More from Meta Global AI Developer Hackathon

- [Proofline](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/111)
- [Cengkudu 800 MDPL](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/112)
- [Medina](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/113)
- [Reality Compiler](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/115)
- [Yebo](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/116)
- [Cleanroom](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/117)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/114. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt