Claude Memory Skill

claude-memory-skill

╔══●══●══●══════════════════════════════════════════════╗
║                                                       ║
║  > tree ~/.claude/memory                              ║
║                                                       ║
║    ├── core.md          claude-memory-skill           ║
║    ├── topics/          ───────────────────           ║
║    │   └── *.md         a skill is all you need       ║
║    └── me.md                                          ║
║                                                       ║
╚═══════════════════════════════════════════════════════╝

*An embarrassingly simple and minimal implementation for agentic memory.*

No databases. No embeddings. No semantic search. Just markdown files and a skill that teaches Claude when to read and write.

The Problem

Claude Code forgets everything between sessions. Built-in auto-memory exists but:

• It's opaque (Claude decides what's "meaningful")
• Limited to 200 lines loaded at startup
• Not tightly integrated into the agentic loop
• No hierarchical organization (scales poorly)

The Solution

A skill-based memory protocol with:

• Hierarchical storage: core.md summaries → topics/.md details
• Background agents: Memory ops don't block the main agent
• Categorized entries: No dumping ground, everything has a topic
• Filesystem-based: Robust, inspectable, git-trackable

Installation

curl -fsSL https://raw.githubusercontent.com/hanfang/claude-memory-skill/main/install.sh | bash

Or clone and run locally:

git clone https://github.com/hanfang/claude-memory-skill.git
cd claude-memory-skill
./install.sh

What Gets Installed

~/.claude/
├── CLAUDE.md              # Hook added (or created)
├── commands/
│   └── mem.md             # The memory skill
└── memory/
    ├── core.md            # Summaries + pointers (always loaded)
    ├── me.md              # About you (always loaded)

...