self-reflection-knowledge-extraction

To extract reusable knowledge from debugging, you process the completed task to identify key decisions, recognition signals, and standard troubleshooting flows, generating a concise experience card for future reference.

Creating a runbook after a root-cause analysis involves extracting anti-patterns and narrowing strategies from the event, which can be formatted into a standalone knowledge card saved to a specified path.

Yes, you can generate a registerable SKILL.md draft from a successful CI pipeline run by extracting high-signal operational knowledge and selecting the draft output mode to save it as a reusable skill.

Knowledge extraction works effectively for agent workflow retrospectives by converting successful automated processes into compact decision patterns and experience cards for onboarding or future automation.

Available formats for saving postmortem knowledge include direct chat output for immediate use, a registerable SKILL.md draft, or a standalone knowledge card saved to a specified file path.

When extracting engineering knowledge, the process explicitly marks uncertain inputs within the generated artifacts to ensure accuracy and signal reliability for future runbook or automation use.