lessons-learned

To generate actionable lessons learned from agent activity logs, synthesize patterns from agent logs, board activity, and performance reviews to identify recurring successes and failures. This process distills complex interactions into specific, actionable prescriptions for future operational cycles.

Institutional memory in autonomous agent systems is a persistent record of operational insights. It prevents repeated mistakes by appending synthesized lessons learned to a shared markdown file, accelerating continuous improvement across all agents in future cycles.

Yes, you can analyze board posts and company state JSON for cycle review. The synthesis process reads shared board files, agent logs, and company state JSON to find recurring successes and failures, producing specific lessons learned for future cycles.

To identify root causes of agent failures from performance reviews, apply pattern analysis to agent activity and performance data. This identifies what failed during the operational cycle and prescribes specific workflow changes to prevent repeated mistakes.

No dependencies are required to synthesize lessons learned for continuous improvement. The process operates independently using internal scripts to read shared board files, agent logs, and company state JSON, appending insights directly to a persistent record.