Workflow State Machines

Overview

This skill provides a structured framework for building resilient, predictable agentic workflows using directed graph-based state machines. It emphasizes durable execution, explicit state transitions, and auditability to ensure complex agent interactions remain stable and debuggable.

When To Use

• When orchestrating multi-step, multi-agent processes that require durability and fault tolerance.
• When you need to integrate human-in-the-loop (HITL) checkpoints into long-running tasks.
• When debugging non-deterministic agent behavior through state history inspection.
• When enforcing KILO-KIT C4 Runtime Protocol stages as formal state transitions.

Core Concepts

Directed Graph Topologies

• Nodes: Represent distinct agent actions, tool executions, or decision points.
• Edges: Define the transition logic based on input, output, or conditions.
• Immutable State: Workflow state is treated as immutable, with transitions resulting in new state snapshots to ensure clean history.