Agent Behavior Tree Stack
Build agent control flow on the BT formalism rather than free-form ReAct, with structured persona configuration and evidence-driven prompt selection.
Description
An agent platform organised around behavior trees from game and robotics AI: leaves are LLM calls or tool invocations returning success/failure; selectors and sequences orchestrate control flow with first-class retry and fallback semantics; trees are constructed by back-chaining from goal post-conditions; agent identity lives as a structured persona profile; prompts are selected via measured variant evaluation.
Open the full interactive page →
Diagram, neighbourhood map, code examples, related patterns and full provenance.