project-brain plugin
A Claude Code plugin: a brain folder structure and handoff protocol so a new session resumes a project without losing context
- 174 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The plugin creates and updates Markdown files in the `brain/` folder of your project
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/project-brainDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The plugin creates and updates Markdown files in the `brain/` folder of your project
This is third-party code. Review the repository files before installing.
What it does
project-brain adds a `brain/` folder of five core files to a project, split by how often they change. `PROJECT.md` describes the essence of the project and what it explicitly does not do, `MAP.md` is a module map and document index, `STATUS.md` holds current state (a soft cap of 80 lines), `DECISIONS.md` keeps a decision log with mandatory rejected alternatives and `HANDOFF.md` carries a shift report for the next window. Past reports sit in `handoffs/` and per-topic notes in `topics/`. The plugin adds a skill with five workflows: new-project kick-off, resume, window-switch handoff (works without `brain/` too), updates with approval of each item and handoff between parallel workstreams. The skill activates only on an explicit request: "set up project brain", "resume this project", "hand off", "update the project brain".
Who it is for. People running a long project with an AI assistant across many sessions and windows who are tired of explaining each time what is going on.
Good fit when
- The window context is running out and a report for the next session is needed before switching
- Decisions drown in commits and chat, and later you need to understand why this decision was made
- One README mixes description, status, history and run instructions
- Parallel workstreams run on a project and confuse who is doing what
Not a fit when
- A one-evening project: a five-file structure costs more than it gives
- Automatic memory without a human is needed: the skill proposes updates and you approve them
- You need a team product with access and roles: the method is local and rests on files
Example request
I am switching windows, prepare a shift handoff for the next session.Limitations
This is a method and a set of templates, with no automatic memory: the structure must be maintained, the skill proposes updates with reasons and you approve each item. The skill activates only on an explicit request and does not trigger just because a `brain/` folder exists. The main installation path targets Claude Code, and for other assistants the structure is copied by hand. The repository includes an example `examples/small-saas`. The latest commit in the repository is from October 2026.
How to disable. Run `/plugin uninstall project-brain@sprout-labs` and `/plugin marketplace remove Ethan-YS/project-brain`. The `brain/` folder in your project stays as ordinary Markdown files.
FAQ
How do I install project-brain?
In Claude Code run `/plugin marketplace add Ethan-YS/project-brain` and `/plugin install project-brain@sprout-labs`. In a terminal the same is done with `claude plugin ...` commands.
Which phrases trigger the skill?
"set up project brain" creates the structure, "resume this project" loads the map and status, "hand off" writes a shift report and "update the project brain" proposes updates with reasons.
Why split files by volatility?
Rare and frequent data in one file interfere: you read a lot and find the needed part with difficulty. Splitting lets you read two or three files and get to work.
Can it be used without Claude Code?
Yes, the structure and protocol are ordinary Markdown files. Copy the `brain/` template and ask the assistant to read `MAP.md` and `STATUS.md` at the start of a session.
Is there an example?
Yes, `examples/small-saas` holds a filled `brain/` folder of a small SaaS: PROJECT, MAP, DECISIONS and HANDOFF.
Discussion
No questions or reviews yet.
Related
PluginsA skill that keeps project memory in plain Markdown files: session logs, a wiki, rules, a to-do list and a note on where to resume
A Claude Code plugin that keeps memory across sessions in one local file, with no database, cloud or API keys
A Claude Code skill collects product context in Markdown and links decisions to interviews, documents and hypotheses
A scaffold for Claude Code and Codex: tasks in files, developer and QA roles, manual approval and shared state.
A set of 58 commands for product managers in Claude Code, Codex and Cursor: from an idea and a PRD to launch and growth analysis
ALIVE gives Claude Code a memory in plain folders on your own computer so every session picks up where the last one left off
README in short
The README describes the problem of long projects with an assistant, then shows the `brain/` folder structure, the principle of splitting by change frequency and plugin installation. The skill workflows and examples are listed separately.