ooda-loop-state-machine-skill

An OODA loop for AI agents organizes decision-making into Observe, Orient, Decide, and Act phases, creating bounded, traceable actions. It enforces a mandatory ooda-cycle-log.md to maintain tempo and accuracy in rapidly changing or adversarial environments.

A state-machine protocol for agent decision-making structures the workflow into five states with explicit exit conditions and gating. This enforces diagnostic artifacts and loop-back re-observation to ensure actions remain bounded and traceable.

Start the OODA cycle by initializing the ooda-cycle-log.md file before the first action. Then begin the Observe phase to start the loop, updating this log each cycle to ensure traceability.

Yes, the OODA loop skill supports safety gating by enforcing mandatory artifacts and gating conditions across its five states. This ensures actions remain bounded and traceable before proceeding through the loop.

You need a structured OODA framework when your agent operates in dynamic or adversarial environments requiring fast, repeatable decision loops. It maintains tempo and accuracy by translating the OODA framework into a state-machine protocol with loop-back re-observation.

The OODA state machine requires creating and updating an ooda-cycle-log.md file before the first action, adding process overhead. It relies on explicit exit conditions and gating, which may reduce tempo if diagnostic artifacts are not managed efficiently.