> For the complete documentation index, see [llms.txt](https://ggai.gitbook.io/ggai-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ggai.gitbook.io/ggai-docs/high-level-architecture/off-chain-ai-engine/ai-modules.md).

# AI Modules

1. **Exploit or Attack Database:** A library of known vulnerabilities or malicious behaviors relevant to Solana programs—e.g., re-initialization vulnerabilities, rent exemption edge cases, or concurrency race conditions.
2. **Machine Learning:** Our model attempt to find novel attacks based on patterns in existing code.
3. **Fuzzer:** Automated fuzz testing to systematically inject random or boundary inputs into program instruction data to see if it causes unexpected behavior.

#### Scaling

As usage grows, the AI engine will scale horizontally by spinning up additional worker nodes. Coordination is handled through a job queue on Redis. Each worker node fetches `ChaosRequest` IDs, performs the chaos simulation, and returns results.