Claude Code Memory Management (CCM)

Ever wished Claude Code remembered your previous conversations?

CCM automatically saves every session, loads context when you return, and helps you search past conversations—so you never lose progress or repeat yourself.

💭 What Problem Does This Solve?

Without CCM:

• 😞 Each new session starts from scratch—Claude has no memory of previous work
• 🔍 Can't search through past conversations to find what you discussed
• 📝 Need to manually copy important decisions or code snippets before exiting
• 🔄 Repeat explanations every time you start a new session

With CCM:

• ✅ Automatic session continuity—previous context loads on startup
• 🔍 Search your entire conversation history across all projects
• 💾 Zero effort—everything saved automatically when you exit
• 🎯 Context-aware—each project has its own conversation history

Perfect for:

• Long-running projects with multiple sessions
• Teams wanting to track decision history
• Developers switching between multiple projects
• Anyone who values conversation continuity

🚀 Quick Start

Installation

# Clone the repository
git clone https://github.com/rexzhen/ccm ~/ccm-plugin

# Add as a marketplace
claude plugin marketplace add ~/ccm-plugin

# Install the plugin
claude plugin install ccm

That's it! The plugin is now active and will automatically manage your sessions.

Usage

Automatic operation (no commands needed):

1. Start Claude Code → Last session summary loads automatically
2. Have your conversation → Transcript captured automatically
3. Exit Claude Code → Session saved automatically with cleanup

Manual commands:

/ccm-save "Custom message"    # Manually save session
/ccm-history                   # Browse recent sessions
/ccm-history "search query"    # Search past sessions

📖 Documentation

• How It Works - Architecture, storage behavior, and project detection
• Usage Examples - Real-world scenarios and workflows
• FAQ - Common questions and troubleshooting

💡 Key Concepts

Automatic Context Detection

CCM automatically detects your working environment:

First session in new folder → Uses ~/.claude/sessions/ (global)
After .claude created       → Switches to <project>/.claude/sessions/
                             (automatic, seamless transition)

Why this matters:

• Works immediately without setup
• Seamlessly transitions when project is initialized
• Each context has independent 500 MB storage limit
• Zero configuration needed

Storage Management

• Default limit: 500 MB per context (project or global)
• Auto-cleanup: Runs on session exit, deletes oldest sessions first
• Configurable: Edit .claude-plugin/config.json to adjust limit
• Zero maintenance: Completely automatic

Learn more about storage management →

🎯 Common Scenarios

Working on Multiple Projects

cd ~/project-a && claude  # Sessions in project-a/.claude/sessions/
cd ~/project-b && claude  # Sessions in project-b/.claude/sessions/

Each project maintains its own independent session history.

Quick Questions Outside Projects

cd ~ && claude  # Sessions in ~/.claude/sessions/ (global)

Global sessions don't clutter project history.

First Time in a New Directory

Day 1: No .claude folder yet → Uses global sessions Day 2: .claude created automatically → Switches to project sessions

See more examples →

🛠️ Configuration

Edit .claude-plugin/config.json to customize:

{
  "maxStorageMB": 500,           // Storage limit in MB
  "cleanupEnabled": true,        // Auto-cleanup on/off
  "preserveLatestSummary": true  // Always keep latest.md
}

Common configurations:

• Conservative: "maxStorageMB": 100
• Generous: "maxStorageMB": 1000
• Disable cleanup: "cleanupEnabled": false (not recommended)

🤖 AI-Powered Summaries (Optional)

CCM can generate intelligent session summaries using Claude AI. This requires the Claude CLI and API credentials.

Setup Options

Option 1: Anthropic API (Direct)

# Install Claude CLI
npm install -g @anthropic-ai/claude-cli

# Set API key in your shell profile (~/.zshrc or ~/.bashrc)
export ANTHROPIC_API_KEY="sk-ant-..."

Option 2: AWS Bedrock

# Install Claude CLI
npm install -g @anthropic-ai/claude-cli

# Configure AWS credentials in your shell profile
export AWS_REGION="us-east-1"
export AWS_ACCESS_KEY_ID="your-access-key"
export AWS_SECRET_ACCESS_KEY="your-secret-key"

# Or use AWS profile
export AWS_PROFILE="your-profile"
export AWS_REGION="us-east-1"

Without AI Setup:

• CCM still works fully with basic pattern-based summaries
• You'll see a one-time notice about Claude CLI setup

Benefits of AI Summaries

• Context-aware understanding of conversations
• Intelligent extraction of key decisions and outcomes
• Deduplication across sessions
• Natural language summaries instead of text fragments

📂 Directory Structure

Project Sessions

your-project/
├── .claude/
│   └── sessions/
│       ├── 2026-01-31T14-30-22-000Z.jsonl
│       ├── summaries/
│       │   └── latest.md
│       └── archives/
└── .gitignore  # Add .claude/sessions/ here

Global Sessions

~/.claude/
└── sessions/
    ├── 2026-01-31T14-30-22-000Z.jsonl
    ├── summaries/
    │   └── latest.md
    └── archives/

🔧 Troubleshooting

Sessions not loading?

/ccm-history  # Check current context and session location

Want project-specific sessions?

# Create .claude directory
mkdir .claude
# Or use Claude Code's init
/init

Storage issues?

# Check config
cat ~/.claude/plugins/ccm/.claude-plugin/config.json

# Adjust storage limit (e.g., to 1 GB)
# Edit config.json: "maxStorageMB": 1000

More troubleshooting →

🗑️ Uninstallation

# Uninstall plugin
claude plugin uninstall ccm

# Optional: Remove session history
rm -rf ~/.claude/sessions/              # Global sessions
rm -rf <project>/.claude/sessions/      # Project sessions

Note: Uninstalling the plugin does NOT delete your session history. Sessions remain on disk unless explicitly removed.

Complete uninstall guide →

📚 Additional Resources

• FAQ - Frequently asked questions
• How It Works - Technical details and architecture
• Examples - Real-world usage scenarios
• GitHub Issues - Report bugs or request features

🤝 Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📄 License

MIT License - see LICENSE file for details

👤 Author

Rex Zhen

• GitHub: @rexzhen
• Email: rex.zhen@gmail.com

---

⭐ If you find CCM useful, please star the repository!

Report an Issue | Request a Feature