The Brain Behind AI Workflow Orchestration
expressiveness(DAG) + rigor(FSM) = reliability. Parallel execution, automatic retry, deterministic state machine.
R(t) = base × e^(-t/stability). Three-tier memory with Ebbinghaus forgetting curve for natural memory decay.
Auto-classify tasks: quick (~2s) → deep (~30s) → visual (~30s) → ultrabrain (~120s). Route to appropriate agents.
pattern → detect → package → skill → MCP tool. Automatic pattern detection and reusable skill packaging.
ETCLOVG Coverage
Tests Passing
Dependencies
Lines of Code
Clone the repository and start building reliable AI agent workflows in minutes.
View on GitHub Read the Docs