ce-session-extract

To extract session skeletons from Claude Code or Cursor JSONL files, you can use this skill's built-in parsing to isolate clean user and assistant turns without loading the full transcript into context.

Finding error signals in AI session transcripts is done by extracting tool failures or command errors for rapid triage, allowing you to isolate concise and actionable insights from large session dumps.

This session extraction tool does not require external dependencies; it only requires Python 3 installed in your environment to run its built-in parsing scripts on your JSONL inputs.

Yes, you can analyze a single Codex session file for a targeted deep-dive, as the skill supports per-session analysis by invoking it with a specific session file path and mode.

You can limit the output size when parsing large AI transcript dumps by using the skill's optional limits feature, which produces structured outputs while keeping the extracted skeleton or error summary concise.