Find skill fixes in your agent conversation history.
Evaluates recent local agent conversations to produce a graded report and evidence-backed drafts of skill improvements.
When to use it
Use when the user wants their agent setup graded from real conversation history or wants to know which installed skills are working.
Ask it to evaluate your agent setup; it asks which conversations and skills to include, then produces a local report and proposed skill edits.
Requires Python 3.9 or newer; the collector uses only the standard library.
This skill
Claude Code
Reads Claude Code conversation history
Codex
Reads Codex conversation history
Warp
Reads Warp conversation databases