What problem does it solve?
Post-campaign learning extractor reads a completed campaign file, its postmortem, and telemetry audit log to extract successful patterns, failed patterns, key decisions, and quality rule candidates. It writes findings to the knowledge base and optionally appends quality rules to harness.json to enable faster, repeatable improvements across future campaigns.
Core Features & Use Cases
- Pattern extraction: reads campaign artifacts to identify successful and failed patterns with concrete evidence.
- Decision capture: aggregates key decisions and rationales to guide future engineering and QA efforts.
- Knowledge write-back: writes results to .planning/knowledge/ and can auto-update harness.json with quality-rule candidates.
- Use case: after a completed campaign, run /learn to surface lessons and enforce repeatable improvements.
Quick Start
Trigger the learning pass by invoking /learn after a campaign completes to generate and persist pattern insights.