Lesson 9: claude-mem Advanced Skills (Part 2) and Practical Workflows

This lesson introduces the remaining five skills of claude-mem and demonstrates how to string them together into real-world development workflows.

9.1 Remaining Core Skills

Skill Name	Core Purpose
`pathfinder`	Architectural Audit. Draws functional flowcharts, identifies cross-cutting concerns, and provides refactoring suggestions.
`knowledge-agent`	Specialized Knowledge Brain. Distills all historical experience regarding a specific topic (e.g., auth, devops) into a conversational sub-agent.
`timeline-report`	Project Retrospective. Concatenates all Observations within a specific timeframe into a narrative progress report.
`how-it-works`	Self-Description. Explains claude-mem's own operational mechanisms and current configuration.
`version-bump`	Automated Release (Mainly for plugin authors). Handles NPM and GitHub release workflows.

Scenario: You are taking over a Python project with 200+ files and plan to refactor its authentication module.

Recommended Workflow:

Build Understanding: Say "Help me understand the overall structure of this project." The LLM automatically calls smart-explore to return an architectural sketch.
Deep Scan: Run learn-codebase specifically on the auth subdirectory.
Architectural Audit: Run pathfinder. It will draw a Mermaid flowchart and point out: "These 3 locations are performing permission checks but using different middleware."
Historical Retrieval: Ask "How did we unify FastAPI middleware in other projects?" This triggers mem-search to find cross-project experience.
Formulate Plan: Run make-plan to generate a refactoring strategy.
Automated Execution: Run do to begin the phased refactoring.

If you need to report progress to your team:

Operation: "Generate a Timeline report for this project for Q2."
Effect: claude-mem pulls all bugfix, feature, and decision observations from the past three months from SQLite and strings them into a Markdown report, showcasing the project's evolution path.