One-click install
npx skills add https://github.com/JimmyMcBride/brain --skill brain-jimmymcbride
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: brain
Source: https://github.com/JimmyMcBride/brain/tree/main/skills/brain
Command: npx skills add https://github.com/JimmyMcBride/brain --skill brain-jimmymcbride

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

brain prevents AI coding agents from repeatedly relearning a repository’s context by turning the project itself into durable, agent-readable memory.

Core Features & Use Cases

  • Project-local memory and retrieval: Keeps durable markdown knowledge in the repo and supports local retrieval over a project-local SQLite index.
  • Compiled startup context with sessions: Prepares validated sessions and compiles task-appropriate context packets with budget control and reuse.
  • Safe, auditable markdown updates: Supports note history and undo for Brain-managed markdown so updates remain traceable as the codebase evolves.
  • Use Case: When switching tasks across files or sessions, use brain to retrieve and compile the correct project context (contracts, workflows, and current state) without wasting turns rediscovering what the repo already knows.

Quick Start

Use brain to prepare the repository’s durable context for your next coding task by running: brain prep --task "<task>".

Frequently Asked Questions about brain

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

FAQPage Schema
How do I maintain durable project memory for AI coding agents across sessions?

Durable project memory is maintained by keeping markdown knowledge directly in your repo and using a local SQLite index for retrieval. This prevents agents from repeatedly relearning repository context during task switching.

What's the best way to compile startup context for an AI coding task?

You compile startup context by running `brain prep --task "<task>"` to generate task-appropriate context packets. This prepares validated sessions with budget control to avoid wasting turns rediscovering existing repo state.

Can I track markdown note history and undo updates as my codebase evolves?

Yes, you can track markdown note history and undo updates. Brain-managed markdown files support safe, auditable updates, ensuring that all knowledge edits remain traceable as your project changes over time.

How do agents retrieve local repository contracts and workflows without wasting tokens?

Agents retrieve local repository contracts by querying a project-local SQLite index. This compiled retrieval mechanism supplies the exact context needed without wasting tokens rediscovering what the repo already knows.

Does this project memory tool require specific repo-local context files to work?

Yes, it requires repo-local Brain docs like AGENTS.md and .brain context to function correctly. The CLI uses these local files first, only falling back to generic behavior when repo-local context is completely absent.

Why does my AI agent keep relearning repository context every time I switch files?

Agents relearn context because they lack durable project memory. Brain solves this by turning the project itself into agent-readable memory, allowing immediate retrieval of contracts, workflows, and current state across files.