agent-brain

Agent ReAct framework with cognitive planning engine — five-phase cognitive cycle with nested ReAct loops, dynamic skill acquisition, and interactive user input.

Cranibox — anytime, anywhere, every device: a self-hosted, trusted Agent Brain. Run Codex, Claude Code & Hermes in your own container with a real desktop, browser and code-server, reachable from any browser or via Tailscale private HTTPS. One /config volume, loopback-only CDP, emergency stop.

Layer 1 projector: RKC OKF → Agent Brain (Chroma + BM25 + Kuzu). Owns no nouns.

Local-first brain repo and trust console for AI agents. Turns local documents into source-backed wiki, graph, memory, claims, and evidence.

Lightweight, domain-agnostic Agent Brain Core — the memory layer for AI agents. Features: entity graph, dream cycle, merge audit, full-text search, web UI.

Evidence-backed Founder OS and governed Agent Brain for capability selection, implementation planning, and supervised AI-agent workflows.