What does the Skill Extractor do?

The Skill Extractor parses Claude Code JSONL conversation logs and utilizes Fabric AI patterns to distill complex debugging sessions into clean, structured, and reusable skills formatted as SKILL.md files.

When should I use this skill?

Use this skill immediately after a successful troubleshooting session, feature implementation, or research conversation to capture the 'working recipe' before the context is lost or buried in history.

How does it improve my developer workflow?

It automates the documentation of tribal knowledge. By converting raw logs into procedural skills, it ensures you and your team can replicate successful solutions without repeating the original trial-and-error process.

What external tools are required for this skill to work?

This skill relies on Fabric AI for pattern extraction and wisdom distillation, and jq for parsing the JSONL conversation files generated by Claude Code.

Does it include failed attempts in the final output?

No. A core principle of the Skill Extractor is to remove noise. It systematically filters out trial-and-error, failed commands, and backtracking to provide a streamlined, functional solution path.

Skill Extractor

Name: Skill Extractor
Author: rafaelcalleja

byrafaelcalleja

•

学習とドキュメンテーション

Transforms conversation history and debugging sessions into structured, reusable skills using AI-driven pattern extraction.

概要

Skill Extractor automates the process of turning troubleshooting history and trial-and-error sessions into professional, actionable documentation. By parsing Claude Code's internal JSONL logs and applying Fabric AI patterns, it identifies the winning recipe from complex debugging journeys, stripping away dead ends to create clean procedural guidance. It serves as a meta-tool for developers looking to preserve tribal knowledge, document complex feature implementations, and scale their technical expertise across projects by converting exploratory conversations into permanent, shareable skills.

主な機能

1 GitHub stars
Generates structured SKILL.md files following progressive disclosure best practices
Distills procedural recipes by systematically removing trial-and-error and non-functional attempts
Parses Claude Code JSONL files into clean transcripts without formatting noise
Utilizes parallel Fabric AI patterns to distill wisdom, instructions, and solutions simultaneously
Extracts core problem patterns and working solutions from raw conversation history

ユースケース

Capturing institutional knowledge from exploratory research conversations
Converting a successful debugging session into a reusable troubleshooting guide
Documenting complex feature implementation workflows discovered during problem-solving

概要

主な機能

1 GitHub stars
Generates structured SKILL.md files following progressive disclosure best practices
Distills procedural recipes by systematically removing trial-and-error and non-functional attempts
Parses Claude Code JSONL files into clean transcripts without formatting noise
Utilizes parallel Fabric AI patterns to distill wisdom, instructions, and solutions simultaneously
Extracts core problem patterns and working solutions from raw conversation history

ユースケース

Capturing institutional knowledge from exploratory research conversations
Converting a successful debugging session into a reusable troubleshooting guide
Documenting complex feature implementation workflows discovered during problem-solving