MIT-licensed agent meta-harness for Claude Code / Codex — swarm coordination, self-learning vector memory, zero-trust federation, and a MetaHarness audit — installed via `npx ruflo init` or the Claude Code plugin path.
DSH integration
Ecosystem-related
Author-claimed
Safety audit
Unaudited
Last verified
2026-08-22
License
MIT
01What can it help you accomplish?
Coordinate multi-agent swarms and autonomous workflows from Claude Code or the CLI
Self-organizing agent swarms (hierarchical / mesh / adaptive topologies), persistent HNSW vector memory, 100+ specialized agents, and a federation layer for cross-machine collaboration
Engineering teams and solo developers running Claude Code / Codex who want agent coordination, memory, and background automation beyond a single model
Audit and harden an agent setup before shipping
A MetaHarness audit that grades agent-setup readiness (1-100), scans tool configs for security issues, and snapshots the project to catch regressions
Teams that need governance / security review of their agent tooling before production use
02How to install into DeepSeek Harness
Prerequisites
- Node.js + npm (the CLI is distributed via `npx ruflo` and `npm install -g ruflo`)
- Claude Code (optional) — for the Claude Code plugin path and MCP server registration
Installation steps
- 01
Install via the one-line POSIX installer: `curl -fsSL https://cdn.jsdelivr.net/gh/ruvnet/ruflo@main/scripts/install.sh | bash`
$ curl -fsSL https://cdn.jsdelivr.net/gh/ruvnet/ruflo@main/scripts/install.sh | bash
- 02
Or run the interactive setup wizard (all platforms, incl. native Windows PowerShell): `npx ruflo@latest init wizard`
$ npx ruflo@latest init wizard
- 03
For Claude Code only: add the marketplace (`/plugin marketplace add ruvnet/ruflo`) then `/plugin install ruflo-core@ruflo`
Verify the integration
- Run `ruflo verify` to cryptographically prove your installed bytes match the signed witness
03DSH integration and capability boundaries
Listed in the DSH plugin catalog (topic `dsh-plugin`); the README documents Ruflo as an agent meta-harness for Claude Code / Codex with a 100+ agent swarm, MCP server, self-learning memory, and zero-trust federation. DeepSeek Harness (`dsh`) specifics are not described in this README.
Swarm coordination & 100+ agents
coding / testing / security / docs / architecture tasks→self-organizing agent swarms in hierarchical, mesh, or adaptive topologies with consensus
Self-learning vector memory (AgentDB + HNSW)
agent trajectories, documents, and knowledge→persistent HNSW-indexed vector memory with sub-ms retrieval across sessions
memory is persisted on disk and retained across sessionsAgent federation (zero-trust)
tasks and work requests across machines / orgs→encrypted cross-installation agent collaboration with mTLS+ed25519 identity and PII stripping
outbound messages are scanned and PII is stripped before leaving your node
04Who is it for? When not to use it?
Good for
- Engineering teams and solo developers running Claude Code / Codex who want agent coordination, memory, and background automation beyond a single model
- Teams that need governance / security review of their agent tooling before production use
Not for
- Federation and multi-provider model routing send data to remote endpoints; federation strips PII and signs messages before they leave your node, but cross-machine collaboration still requires network access and trust configuration.
05Compatibility, maintenance and safety notes
- The one-line `curl ... | bash` installer needs a POSIX shell (Git-Bash, WSL, MSYS); on native Windows PowerShell / cmd use `npx ruflo@latest init wizard` instead.
- Federation and multi-provider model routing send data to remote endpoints; federation strips PII and signs messages before they leave your node, but cross-machine collaboration still requires network access and trust configuration.
- Multi-provider model routing (Claude, GPT, Gemini, Cohere, Ollama) requires an API key / endpoint for each provider you enable; the hosted Web UI uses OpenRouter.
MIT · actively maintained (latest release v3.38.16, 2026-08-21)
06Frequently asked questions
How is Ruflo related to DeepSeek Harness (dsh)?
Ruflo is listed in the DSH plugin catalog as a `dsh-plugin`, but this README only documents its Claude Code / Codex integration (100+ agents, MCP server, swarms). To use it under dsh you would bridge it via its MCP-server registration (`claude mcp add ...`-style); the README gives no dsh-specific commands.
How do I install Ruflo?
On POSIX: `curl -fsSL https://cdn.jsdelivr.net/gh/ruvnet/ruflo@main/scripts/install.sh | bash`. On any platform including native Windows PowerShell: `npx ruflo@latest init wizard`. For Claude Code only, first `/plugin marketplace add ruvnet/ruflo`, then `/plugin install ruflo-core@ruflo`.
How do I verify the install?
Run `ruflo verify` to cryptographically prove your installed bytes match the signed witness — trust but verify.
Is memory and federation data safe?
Vector memory (AgentDB + HNSW) is persisted to disk and retained across sessions. Federation strips PII (a 14-type detection pipeline) and verifies identity with mTLS + ed25519 before any message leaves your node; untrusted agents see only discovery info, not your memory.
07Related DSH workflows
reactive-resume
by reactive-resume
A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source and free forever. Try it out today!
everos
by evermind-ai
One portable memory layer for every AI agent: local-first, Markdown-native, user-owned, and self-evolving across apps, tools, and workflows.
yao
by yaoapp
✨ All your agents and workspaces in one place, on every device you own. Track tasks on a board, accessible from desktop, mobile, browser, or API. Self-hosted.
openpencil
by zseven-w
The world's first open-source AI-native vector design tool and the first to feature concurrent Agent Teams. Design-as-Code. Turn prompts into UI directly on the live canvas. A modern alternative to Pencil.
08Data and sources
An agent meta-harness for Claude Code and Codex.
One `npx ruflo init` gives Claude Code a nervous system: agents self-organize into swarms, learn from every task, rememb…
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-08-22. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.