Get the user's Ryu Core node running and serving a local model so this plugin's ryu_* tools work. Covers what Ryu is, launching Core or the desktop app, verifying health on 127.0.0.1:7980, and activating a local model. The plugin already wires the MCP bridge, so this skips manual MCP config. Use when ryu_health fails or the user asks to install, start, or connect to Ryu.
Scanned 9/2/2026
Install to Claude Code
npx -y skills add amajorai/ryu --skill setup-ryu --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Setup Ryu?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/amajorai-setup-ryu)More formats (shields.io, HTML) on the badges page.
---
name: setup-ryu
description: Get the user's Ryu Core node running and serving a local model so this plugin's ryu_* tools work. Covers what Ryu is, launching Core or the desktop app, verifying health on 127.0.0.1:7980, and activating a local model. The plugin already wires the MCP bridge, so this skips manual MCP config. Use when ryu_health fails or the user asks to install, start, or connect to Ryu.
---
# Set up Ryu (plugin edition)
This plugin already wires the Ryu MCP bridge into Claude Code, so you do not configure `mcpServers` by hand. Your job is to make sure the node this plugin points at is actually running and serving a model. Work top to bottom and do not skip the health check.
## What Ryu is
Ryu is a local-first AI node on the user's own machine. Two pieces matter here:
- Core - the brain and HTTP API on `http://127.0.0.1:7980`. Owns the model catalog, agents, workflows, RAG Spaces, MCP registry, and chat. Every `ryu_*` tool in this plugin talks to Core.
- Gateway - the OpenAI-compatible inference gateway on `http://127.0.0.1:7981`. Serves the active local model and meters usage.
The desktop app boots both. A headless user can run just Core.
## Step 1 - make sure Core is up
- Desktop: have the user install and open the Ryu desktop app. It starts Core and Gateway automatically.
- Headless from the repo: Core is a Rust binary under `apps/core`; run it with the project's standard run command. It binds `:7980`.
The plugin's Core URL is set from the `ryu_core_url` plugin option (default `http://127.0.0.1:7980`). If the node needs auth, set `ryu_core_token`.
## Step 2 - verify health (do not skip)
Call the `ryu_health` tool. A healthy node returns `{"status":"ok"}`. If it errors, Core is not reachable: recheck Step 1 and confirm the `ryu_core_url` option matches where Core is actually listening. Then call `ryu_system_status` for engine and sidecar state.
## Step 3 - activate a local model
If `ryu_get_active_model` shows nothing serving, help the user pick one:
1. `ryu_search_models` with a query (for example `qwen3`) to browse the catalog.
2. Install a specific GGUF from the Ryu desktop model store, or via the models REST surface (see the `ryu-local-model` skill).
3. `ryu_set_active_model` with the installed model id, then `ryu_get_active_model` to confirm.
## Step 4 - confirm the loop
Call `ryu_health`, then `ryu_get_active_model`. If both succeed, the plugin is fully wired to a live node and every other `ryu_*` tool and `/ryu:*` command will work.
## Where to go next
- Full tool reference and driving patterns: [[ryu-tools]]
- Search, download, and serve local models in depth: [[ryu-local-model]]
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!