omk-control-loop-debugger

What problem does it solve?

It turns confusing agent failures into a structured feedback-loop diagnosis so you can identify whether the issue is setpoint ambiguity, stale or missing evidence, a wrong control decision, insufficient actuator authority, or a hidden disturbance.

Core Features & Use Cases

• Industrial control-loop framing: Models the agent runtime as setpoint, sensors, controller, actuator, disturbance, and correction to separate root causes from symptoms.
• Failure pattern recognition: Targets oscillation, saturation, sensor drift, bad actuator effects, and controller mismatch (wrong skill/worker/provider/route).
• Evidence-driven correction plan: Produces a verification requirement that demonstrates loop stability rather than guessing at causes.

Quick Start

Use the omk-control-loop-debugger to diagnose why your OMK agent workflow oscillates, stalls, misroutes tools, loses context, or fails evidence-gated decisions by filling in setpoint, sensor evidence, control error, controller decision, actuator path, disturbance, correction, verification, and residual risk.