What happens if I already have a CLAUDE.md file?

The skill intelligently reads your existing file and uses an 'Edit' tool to insert new insights into the most semantically appropriate section rather than just appending to the end.

Does it save everything automatically without my consent?

No. The skill follows a strict user confirmation flow where it presents a summary of findings, allowing you to select exactly which items to persist or skip.

Where does the auto-extract skill save my information?

It persists data to three layers: Project CLAUDE.md for repository rules, User CLAUDE.md for global preferences, and Memory MCP for situational knowledge recall.

How does it decide between project-specific and user-level info?

It uses a logical decision tree: toolchain or architectural info goes to the project layer, identity and high-frequency preferences go to the user layer, and situational experiences go to Memory MCP.

Session Learning Auto-Extract

Name: Session Learning Auto-Extract
Author: cuipengfei

bycuipengfei

•

생산성 및 워크플로

Analyzes AI sessions to automatically extract and persist key insights, project rules, and user preferences across multiple storage layers.

소개

The Session Learning Auto-Extract skill transforms Claude from a stateless assistant into an evolving development partner by capturing valuable session data. It systematically reviews interactions to identify successful solutions, user preferences, and project-specific constraints. By intelligently categorizing information into project-level CLAUDE.md files, global user settings, or the long-term Memory MCP, it ensures that lessons learned in one session are seamlessly carried forward, significantly reducing repetitive instructions and improving contextual accuracy over time.

주요 기능

Intelligent decision tree for automated insight categorization
5 GitHub stars
Structured memory cleanup and graph maintenance checklist
Three-tier storage architecture (Project, User, and Memory MCP)
Interactive user confirmation workflow for data persistence
Seamless merging with existing CLAUDE.md documentation structures

사용 사례

Capturing project-specific architectural decisions and toolchain configurations
Persisting user-specific coding styles and communication preferences across all projects
Documenting complex bug fixes and situational 'gotchas' for future recall

소개

주요 기능

Intelligent decision tree for automated insight categorization
5 GitHub stars
Structured memory cleanup and graph maintenance checklist
Three-tier storage architecture (Project, User, and Memory MCP)
Interactive user confirmation workflow for data persistence
Seamless merging with existing CLAUDE.md documentation structures

사용 사례

Capturing project-specific architectural decisions and toolchain configurations
Persisting user-specific coding styles and communication preferences across all projects
Documenting complex bug fixes and situational 'gotchas' for future recall