Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Add Atomic Chat Local-Model Tool (NanoClaw)

Wires a stdio MCP server into the NanoClaw agent-runner so the containerized agent can call local models served by the Atomic Chat desktop app.

Dev & CodingAdvanced★ 30,840⑂ 12,826AI score 7/10Last updated: Sep 23, 2026

What it does

  • Copies a stdio MCP server (atomic-chat-mcp-stdio.ts) into container/agent-runner/src and registers an atomic_chat entry in the mcpServers map in index.ts.
  • Adds ...atomicChatEnv() to contributedEnv inside composeSessionSpec so ATOMIC_CHAT_HOST / ATOMIC_CHAT_API_KEY are forwarded into the container.
  • Patches the Docker driver's stderr handler so [ATOMIC] lines log at info level, and appends env stubs to .env.example.
  • Exposes two agent tools: atomic_chat_list_models and atomic_chat_generate (OpenAI-compatible calls to 127.0.0.1:1337/v1).
  • Includes vitest/bun wiring tests plus a troubleshooting section for connection, 404, and cold-start issues.

Who it's for

  • Operators of the NanoClaw stack who want Claude to orchestrate while offloading bulk work to local models.
  • Users already running Atomic Chat (a Jan fork) with its Local API Server enabled.
  • macOS + Docker Desktop setups where the container reaches the host via host.docker.internal.

Examples

  1. Say "use atomic chat to tell me the capital of France" — the agent lists models, then generates locally.
  2. "Use atomic_chat_generate with llama3.2-3b-instruct to summarize this log" — heavy text work at zero API cost.
  3. Run tail -f logs/nanoclaw.log | grep -i atomic to watch [ATOMIC] <<< Done: model | Xs | N tokens lines for latency checks.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/nanocoai/nanoclaw/HEAD/.claude/skills/add-atomic-chat-tool/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the .claude/skills/add-atomic-chat-tool folder from the GitHub repo nanocoai/nanoclaw into my ~/.claude/skills/add-atomic-chat-tool/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/nanocoai/nanoclaw.git && cp -r nanoclaw/.claude/skills/add-atomic-chat-tool ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal at your NanoClaw project root (or clone it: git clone https://github.com/nanocoai/nanoclaw.git).
  2. Install the skill: cp -r nanoclaw/.claude/skills/add-atomic-chat-tool ~/.claude/skills/, or just use the copy already in the repo's .claude/skills/.
  3. Install and launch the Atomic Chat desktop app, then enable Settings → Local API Server on port 1337.
  4. Download at least one model from the Hub tab and send one message in the UI to warm it up.
  5. Confirm the API responds: curl -s http://127.0.0.1:1337/v1/models.
  6. In Claude Code, ask it to "apply the add-atomic-chat-tool skill" so it copies files and edits index.ts / container-runner.ts.
  7. Validate: pnpm run build, the provided vitest and bun tests, then ./container/build.sh — all must be clean.
  8. Restart the service (macOS: launchctl kickstart) and verify with a prompt like "use atomic chat to ...".
View source on GitHub ↗License: MIT