Everything here exists for one reason: to make an AI agent's reasoning visible, verifiable, and shareable.
Reasoning streams to the console as the agent forms it. You see the thinking happen — not a polished summary written after the fact.
Every trace is SHA-256 hashed and optionally anchored to Solana. Tamper with one step and the fingerprint breaks. Verifiable by anyone, forever.
Each illumination gets a permalink. Hand someone the proof of how a conclusion was reached, not just the conclusion itself.
Wraps any OpenClaw agent. Keep your skills and tools; Light in Claw illuminates the path they take.
Reasoning rendered in clear, structured language — readable by a human, parseable by an auditor.
Powered by Claude Haiku 4.5 via OpenRouter. Illumination keeps pace with the agent instead of slowing it down.