mycelium-insight

Collect insights from Claude Code sessions and write them to a JSONL file.

Updated May 15, 2026
One-click install
npx skills add https://github.com/heyu-ai/yibi-stack --skill mycelium-insight
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mycelium-insight
Source: https://github.com/heyu-ai/yibi-stack/tree/main/plugins/growth/skills/mycelium/insight
Command: npx skills add https://github.com/heyu-ai/yibi-stack --skill mycelium-insight

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill automatically collects and accumulates insights from Claude Code sessions, ensuring valuable knowledge is preserved for future reference and learning.

Core Features & Use Cases

  • Automated Insight Collection: Stops Claude Code sessions to extract insights in real-time and saves them in a JSONL format.
  • Cross-Agent Integration: Combines metadata from different agents and accounts, facilitating a holistic view of the knowledge collected.
  • Use Case: Imagine you are working on a complex project with multiple team members. This Skill can help capture insights generated by different team members, enabling better collaboration and knowledge sharing.

Quick Start

Run the uv run --directory "$SKILL_REPO" python -m tasks.mycelium insight install-hook command to install the Stop hook.

Frequently Asked Questions about mycelium-insight

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automatically collect insights from Claude Code sessions?

You can automatically collect insights from Claude Code sessions by installing a Stop hook via a Python command, which parses session transcripts in real-time and saves the extracted knowledge to a JSONL file.

What is continuous learning and knowledge preservation in AI coding agents?

Continuous learning and knowledge preservation in AI coding agents involve automatically extracting valuable insights from session transcripts and accumulating them in a structured JSONL file for future reference across multiple agents.

Can I combine session metadata across different Claude Code agents and accounts?

Yes, you can combine session metadata across different Claude Code agents and accounts, as the Skill integrates cross-agent metadata to facilitate a holistic view of the collected knowledge for better team collaboration.

Do I need Python to extract insights from Claude Code session transcripts?

Yes, you need Python to extract insights from Claude Code session transcripts, because the Skill relies on Python scripts to parse the transcripts and write the accumulated insights into a JSONL file.

How does a Stop hook work for real-time session analysis in Claude Code?

A Stop hook for real-time session analysis works by intercepting Claude Code sessions when they stop, automatically triggering a Python script to parse the transcript and extract insights before writing them to a JSONL file.