Where are the memory files stored?

All files are stored locally within your project in the .memory-bank/ directory, allowing you to version control your project's context alongside your code.

How does Memory Bank improve AI performance?

By providing a persistent, structured source of truth, it prevents the AI from losing track of project-specific rules, goals, and architectural patterns between different chat sessions.

Can I use Memory Bank for team collaboration?

Yes, because the memory bank is stored as Markdown files in your repository, it serves as a shared, human-readable context that both developers and AI can maintain.

What core files make up the Memory Bank structure?

The structure includes projectbrief.md, productContext.md, activeContext.md, systemPatterns.md, techContext.md, and progress.md.

When should the memory bank be updated?

Updates should occur whenever new patterns are discovered, after implementing major changes, or when the user explicitly requests a context sync.

Memory Bank

Name: Memory Bank
Author: cuipengfei

bycuipengfei

•

Productivity & Workflow

Manages cross-session project knowledge and context persistence to provide a single source of truth for AI interactions.

Memory Bank is a specialized Claude Code skill designed to overcome session-based memory limitations by maintaining a structured knowledge repository within your project. It establishes a dedicated directory containing core documentation like project briefs, system architecture, and technical constraints, ensuring that the AI has access to accurate context every time a task begins. By actively updating these records after major changes or new discoveries, the skill facilitates a seamless development experience where the AI remains aligned with the project's unique patterns, goals, and evolving decision history.

Key Features

01Automated tracking of project goals, system patterns, and technical stacks

02Proactive maintenance triggers to record decision evolution and resolve ambiguity

035 GitHub stars

04Integrated Plan-Action workflow that validates context before task execution

05Structured context management using a dedicated .memory-bank/ directory

06Single Source of Truth (SSOT) architecture for cross-session consistency

Use Cases

01Maintaining long-term project context across multiple discontinuous development sessions

02Tracking the evolution of features and bug fixes through a persistent record of project progress

03Onboarding AI agents to complex codebases with clear documentation of architectural decisions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cuipengfei/prompts memory-bank

For use in Claude.ai and ChatGPT

Download Skill