zouroboros-introspect

What problem does it solve?

Detects and ranks systemic degradation across Zouroboros subsystems so engineers and autonomous governors can prioritize fixes before failures cascade.

Core Features & Use Cases

• Composite Health Scoring: Aggregates weighted metrics (memory recall, graph connectivity, routing accuracy, eval calibration, procedure freshness, episode velocity) into a 0–100 composite score.
• Per-Metric Diagnostics: Produces per-metric scores, trends, and HEALTHY/WARNING/CRITICAL status with recommended remediation actions.
• Operational Integration: Exports JSON, optionally persists the scorecard as an episodic memory, and feeds the prescription engine for automated evolution.
• Use Case: Schedule a daily 05:00 introspection to detect regressions in multi-agent routing or memory recall and automatically generate prioritized playbook recommendations.

Quick Start

Run the introspection now, produce the ranked scorecard, and store the result as an episodic memory.