memory-map

Configure per-project memory and history storage for Claude Code sessions.

186|15|Updated Apr 2, 2026

Search Tags:#memory #history #claude.md #claude-code #cross-project #session-start #per-project

Ready to run

npx skills add https://github.com/kid-sid/claude-spellbook --skill memory-map

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: memory-map
Source: https://github.com/kid-sid/claude-spellbook/tree/main/skills/memory-map
Command: npx skills add https://github.com/kid-sid/claude-spellbook --skill memory-map

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Memory-map centralizes per-project memory and history management for Claude Code, ensuring session context persists across machines and projects.

Core Features & Use Cases

Install and configure memory_map to control what data is kept in memory vs history
Manage cross-project memory recall and session-start loading via CLAUDE.md
Tune history compression and enable per-project context persistence across repositories

Quick Start

Install memory_map, register the MCP server for the user scope, and copy CLAUDE.md into your project to enable session-start loading.

Frequently Asked Questions about memory-map

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

FAQPage Schema

How do I persist Claude Code session context across projects?▼

Persist Claude Code session context by configuring per-project memory and history files, ensuring data survives across machines and repositories. This Skill manages local storage files to retain what data is kept in memory vs history.

What is the difference between memory and history in Claude Code sessions?▼

Memory stores persistent context for future session loading, while history retains past session data with optional compression tuning. This Skill helps configure what data is kept in memory vs history for per-project use.

How do I set up CLAUDE.md to load memory automatically at session start?▼

Copy CLAUDE.md into your project and register the MCP server for the user scope to enable session-start loading. This triggers automatic memory recall and cross-project context initialization when a session begins.

Can I use per-project memory across multiple repositories?▼

Yes, per-project memory and history can be managed across repositories. The Skill configures local .mcp_memory.json and .mcp_history.json files per project, enabling cross-project memory recall and context persistence.

Do I need to register an MCP server to manage Claude Code history?▼

Optional per-project registration of MCP servers is supported alongside local storage of .mcp_memory.json and .mcp_history.json files. Registration enables cross-project memory workflows, but local files provide the core persistence.

How do I tune history compression for Claude Code sessions?▼

Tune history compression through the memory_map configuration to control how past session data is stored in .mcp_history.json files. This allows you to balance context retention with storage size per project.