Introduction
The short product model for the CHAOS beta docs.
Product model
CHAOS wraps AI coding work in a local operating layer. The goal is to make agent runs less ephemeral by keeping memory, task flow, provider calls, permissions, and audit evidence connected to the repository.
Core surfaces
- Local operator surface for run visibility.
- Context hydration for current repository work.
- PM-orchestrated execution for multi-step tasks.
- MCP-facing integration for external clients.
- Audit records for decisions, prompts, tools, and results.
Beta expectation
The first beta docs should help a tester install the local runtime, connect a provider, run a small workflow, and inspect what happened afterward.