Agent Brain plugin
A plugin with a local document and code index: hybrid semantic and keyword search, a dependency graph and three helper agents
- 120 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Code and documents are sent to the selected cloud embedding and summary providers during indexing
- The plugin starts a local server and stores the project index on disk
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/agent-brainDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Code and documents are sent to the selected cloud embedding and summary providers during indexing
- The plugin starts a local server and stores the project index on disk
This is third-party code. Review the repository files before installing.
What it does
Agent Brain indexes a project's documentation and source code and answers questions in plain language. The Claude Code plugin brings 30 slash commands, three agents and two skills, backed by a FastAPI server and an `agent-brain` command that runs indexing and search. There are five search modes: hybrid by default, vector for concepts, BM25 for error messages and function names, a graph for questions like "what calls X", and a combined mode that merges results. Code is chunked by structure with tree-sitter and supports Python, TypeScript, JavaScript, Java, Kotlin, C, C++, C#, Go, Rust and Swift. Each project folder gets its own server. Embeddings come from OpenAI, Cohere or Ollama and summaries from Anthropic, OpenAI, Gemini, Grok or Ollama. With Ollama everything runs locally. The `/agent-brain-setup` wizard handles first setup.
Who it is for. Developers with a large codebase or documentation who want an agent to find the right place by meaning rather than by exact word.
Good fit when
- A project has hundreds of documentation files and the question "how is authentication built" must be checked across text and code at once
- You need to find which classes use `AuthService`: graph mode shows dependencies
- You need an exact error message or function name, which is what BM25 mode is for
- Sources cannot leave the machine, so everything is indexed locally with Ollama models
Not a fit when
- The repository is small and fits in the agent's context: an index only adds setup
- You need search without Python: the `agent-brain-rag` and `agent-brain-cli` packages install through pip
- Code must not go to cloud embedding providers and Ollama cannot be installed
Example request
Find how authentication is implemented across all documentation and code and show file references.Limitations
The Python packages `agent-brain-rag` and `agent-brain-cli` are required, installed by the setup wizard or pip. Cloud providers need your own OpenAI, Cohere, Anthropic, Gemini or Grok keys, and code and documents go to those services during indexing. A fully local mode requires Ollama. Each project runs its own server on an automatically chosen port, and the index is stored on disk. The README is marked as validated on 16 March 2026, and the last commit is from September 2026.
How to disable. Stop the server with /agent-brain-stop and remove the plugin through /plugin. Delete the project index directory and the agent-brain packages by hand.
FAQ
How do I install the plugin in Claude Code?
Add the marketplace with `claude plugins marketplace add SpillwaveSolutions/agent-brain`, then run `claude plugins install agent-brain@agent-brain-marketplace`. The plugin does not install through pip.
How do I connect it to Codex or Cursor?
Install `pip install agent-brain-rag agent-brain-cli agent-brain-ag-mcp`, then run `agent-brain install-agent --agent codex --with-mcp` or `--agent cursor`. The command converts the plugin to the host's native layout and registers the MCP server.
Which search mode should I choose?
Use `/agent-brain-search` for general questions, `/agent-brain-keyword` for error messages and function names and `/agent-brain-graph` for dependencies. The `using-agent-brain` skill picks a mode on its own.
Can it run without the cloud?
Yes. In `/agent-brain-providers` pick Ollama for both embeddings and summaries. Documents then stay on your machine.
Which agents does the plugin include?
Three: Search Assistant searches across modes, Research Assistant digs into architecture with follow-up queries and Setup Assistant helps with installation and faults.
Discussion
No questions or reviews yet.
Related
PluginsA Claude Code and Codex plugin with a local RuvNet source base: the agent finds the right tool and shows the file with its code.
BrainDB keeps agent memory in PostgreSQL: entities, relations, search and rules, with two skills wiring it to Claude Code.
BrainAPI turns text into a knowledge graph on your server and serves agents over MCP, REST and a console with fact-path answers
AI Builder Club Skills
AI Builder Club — Skills
A Claude Code plugin marketplace to make a repo agent-ready and run compounding agent loops
Claude Code plugin with 60 subagents and 62 team presets, plus a browser designer for planning agent teams visually.
Coding memory across sessions, machines and teammates: decisions and dead ends flow into later prompts on their own.
README in short
The README presents Agent Brain as a document and code search system for agents and apps. It covers plugin installation, the setup wizard, search and server commands, agents and skills, search modes, pluggable providers and a local Ollama mode. The CLI, project structure and technology stack are listed separately.