Action Execution Pipeline With Evaluators And State Composition

TypeScript framework for autonomous AI agents — multi-platform, plugins, memory, social agents.

Unique: Uses state composition pattern to merge character context, memory, and dynamic provider data into unified state object for model decision-making, rather than passing separate context arrays. Evaluators provide pre-execution validation without requiring model calls, enabling efficient constraint checking.

vs others: More structured than LangChain's tool calling (explicit state composition and evaluators) but requires more boilerplate; better for complex decision logic than simple function-calling approaches.