Context Health Monitor

You can detect context rot by tracking signals like repeated debugging, circular reasoning, and uncertainty in AI agent sessions. Monitoring these patterns reveals when decision quality declines and triggers diagnostic state dumps.

Context health monitoring tracks complexity accumulation in AI sessions to protect integrity. It evaluates signals like repeated debugging and uncertainty to detect early signs of context rot and prevent silent quality degradation.

You save an AI session state by triggering automatic state dumps to `.agent/state/STATE.md`. This records diagnostic snapshots when context complexity thresholds are reached, preserving session integrity for recovery.

Yes, you can pause and resume AI sessions to prevent context rot. Guided recovery workflows offer actionable recommendations and safe-guarded options to pause or reset, ensuring a fresh perspective and preserving reliability.

Yes, context monitoring works without external dependencies. The skill operates independently to track context complexity signals and enforce guardrails within AI sessions without requiring additional components or libraries.

You should trigger a state dump when context complexity thresholds are reached in an AI debugging session. Repeated debugging, circular reasoning, and uncertainty logging indicate quality degradation, signaling the need to record a diagnostic snapshot.