RuvNet Brain plugin
A Claude Code and Codex plugin with a local RuvNet source base: the agent finds the right tool and shows the file with its code.
- 83 ★ on the repository
- Last commit
Extended access
Why this level
- The plugins install lifecycle hooks in Claude Code and Codex and download signed artifacts
- Optional executors can run tools through routing if you authorize them
- Project note capture goes into the local AgentDB store with your consent
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/ruvnet-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.
- The plugins install lifecycle hooks in Claude Code and Codex and download signed artifacts
- Optional executors can run tools through routing if you authorize them
- Project note capture goes into the local AgentDB store with your consent
This is third-party code. Review the repository files before installing.
What it does
RuvNet Brain compiles the public source and documentation of Reuven Cohen (rUv) ecosystem into a signed local knowledge bundle. Native plugins for Claude Code and Codex go alongside: the `search_ruvnet` MCP tool, skills and lifecycle hooks. The agent answers your query naming the repository and file and distinguishing a working implementation from a design document. Search uses routing, lexical and vector retrieval and a shared cross-encoder rerank and needs no external embedding key. `/brain-prompt` and `/brain-build` help refine a task and give a bounded build contract. The local RNBC console opens through `/rnbc` and shows installed components and settings, and `ruvnet-brain --doctor` checks the installation. Project note capture happens only with consent and in the project store, and recorded observations are redacted.
Who it is for. Developers adopting RuvNet ecosystem components such as Ruflo, RuVector, AgentDB or SPARC who want to rely on source code rather than retellings.
Good fit when
- You need persistent memory for agents and want to know which RuvNet options fit and where their APIs are implemented
- You must work out how RVF stores and queries vectors, with links to source
- You will compare approaches to routing model calls and want the trade-offs before choosing
- A session resumes after a break and you need to recall what changed and which decisions are confirmed by code or tests
Not a fit when
- The project does not use the RuvNet ecosystem: the knowledge base covers only its public repositories
- You need a guarantee that a capability card tool is really implemented: a card helps find a product and does not prove implementation
- Hooks and a local knowledge bundle cannot be installed: the plugins and artifact downloads are required
Example request
How does Ruflo coordinate agents? Show the source that implements it and separate what is confirmed by code from what is only described in documents.Limitations
Node.js 18 or newer, npm and a supported Claude Code or Codex are required. Installation and updates download dependencies and signed release artifacts. Hook delivery depends on the host, OS and installed tools, so check `ruvnet-brain --doctor` and not only registration. Recommendations do not install tools or authorize spending. Optional task routing needs available and authorized executors. The project covers one author ecosystem and completeness depends on the release census.
How to disable. Disable the plugins in Claude Code or Codex and delete the `~/.cache/ruvnet-brain` directory with the knowledge bundle. `ruvnet-brain --doctor` shows update settings and install state before you remove anything.
FAQ
How do I install RuvNet Brain?
With `npx ruvnet-brain@latest`. The installer downloads the signed knowledge bundle, installs a local reader and configures detected hosts. Then restart Claude Code or Codex and ask about the ecosystem.
Do I need an embedding key?
No. Local retrieval does not use an external embedding API. Your coding host and optional model providers keep their own accounts and billing.
How do I open the console?
Type `/rnbc` or `/rnb` in Claude Code or Codex. A local page opens that scans your environment and explains the effect of each action before you apply it.
How do I check that the install works?
Run `ruvnet-brain --doctor`. Without the command on `PATH` run `npx ruvnet-brain@latest --doctor`. A registered hook alone does not prove delivery, so read the check result.
What can I ask Brain?
A need such as persistent memory for agents, how something works such as RVF vector storage, a comparison of model routing approaches or a request to build a bounded change through `/brain-build`. Ask for the files it read.
Discussion
No questions or reviews yet.
Related
PluginsA plugin with a local document and code index: hybrid semantic and keyword search, a dependency graph and three helper agents
The ak command installs, repairs and checks ruflo and agentic-qe, adds the offline RuvNet Brain base and links Claude Code, Codex
BrainAPI turns text into a knowledge graph on your server and serves agents over MCP, REST and a console with fact-path answers
ReMe keeps agent memory in Markdown files with wikilinks and installs as a plugin for Claude Code, Codex and Qoder
BrainDB keeps agent memory in PostgreSQL: entities, relations, search and rules, with two skills wiring it to Claude Code.
AgentDB keeps agent memory in an .rvf file and, with 41 MCP tools, recalls past solutions while learning from feedback
README in short
The description explains the purpose: helping find and use the RuvNet ecosystem based on source. A one-command install, sample prompts, a three-layer scheme (knowledge, retrieval, action), workflow capabilities, a section on staying current and on evidence caveats follow.