session-retrospective

You can analyze Claude session logs by parsing JSONL files to extract user and assistant exchanges, errors, and corrections, which are then compiled into structured retrospective summaries documenting key challenges and techniques.

Automating retrospective reviews of AI assistant interactions is done by processing session JSON logs to produce detailed markdown summaries capturing challenges, mistakes, and techniques, which are then archived as tracked issues for progress monitoring.

You generate a retrospective summary by invoking the skill with your session ID and target issue key, prompting it to parse the JSONL session log and post the resulting reflective summary directly to the linked issue.

Yes, you can use session review to improve prompt engineering and debugging strategies because the retrospective summaries highlight specific mistakes and techniques from your Claude interactions, facilitating continuous improvement of your AI sessions.

No external dependencies are required for retrospective analysis; the Skill operates independently using its bundled scripts and references to parse JSONL session logs and generate structured markdown retrospectives without external libraries.

When parsing JSONL session data, limitations include reliance on valid session JSON logs for extraction accuracy, and the requirement to provide a target issue key for archiving, meaning incomplete logs may result in fragmented retrospective summaries.