Haiku Evaluator

Haiku Evaluator is a lightweight, native evaluator model developed by Anthropic for Claude Code's /goal feature, checking task completion after every agent step to enable uninterrupted autonomous operation until conditions are met.

Two-hop subgraph: this entity, every entity it directly relates to, and every entity those neighbors relate to. Drag a node, scroll to zoom, click to inspect — or click any neighbor and re-center the atlas there.