> ## Documentation Index
> Fetch the complete documentation index at: https://docs.snapsec.co/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Agent & Testing Engine

> The autonomous agent architecture, tool selection heuristics, and safety boundaries.

Snapsec's AI agent operates on an autonomous ReAct (Reasoning + Action) architecture specifically tuned for offensive cybersecurity.

***

## Architecture Components

* **Offensive LLM Orchestrator**: Evaluates HTTP responses, recognizes error patterns, and synthesizes custom exploit payloads.
* **Sandboxed Toolchain Runtime**: Executes network tools, headless browsers, and scanners in isolated, ephemeral Linux containers.
* **Evidence Collector**: Captures full raw request and response data, eliminating ambiguity.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.