insights

Analyze agent sessions by running retrospective workflows that audit cross-session rule adherence, skill activation, and hook behavior over a weekly window, producing a structured report with diffs to rules, skills, and hooks.

Agent self-improvement via retrospectives is the process of auditing session governance and compliance to surface systemic issues, categorize problems through a phase-driven workflow, and propose actionable fixes to improve agent reliability.

Audit cross-session rule adherence and skill activation by triggering a weekly retrospective that scans agent sessions, diagnoses recurring patterns, and generates prioritized proposals mapped to rules, skills, and hooks for targeted governance improvements.

Yes, you can generate structured diffs to rules, skills, and hooks by running the retrospective workflow which prescribes targeted improvements and outputs a structured report detailing necessary modifications to agent configurations.

Retrospective session analysis works for weekly agent reliability audits by scanning a 7-day window of agent sessions to identify systemic issues, audit compliance, and propose actionable fixes that drive governance improvements.

The best way to surface recurring patterns in agent hook behavior and rule compliance is to execute a phase-driven workflow that scans, audits, diagnoses, and prescribes fixes, categorizing problems into prioritized proposals for targeted improvements.