skills.rest
wrightchin
wrightchinCOMMUNITY
Explore all skills by this author →
GitHub

add-atomic-chat-tool

Expose local Atomic Chat models as MCP tools for NanoClaw agents.

Updated Apr 26, 2026
Search Tags:#integration#mcp#local-models#container-agent#atomic-chat#secure-inference
One-click install
npx skills add https://github.com/wrightchin/nanoclaw-playground --skill add-atomic-chat-tool-wrightchin
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: add-atomic-chat-tool
Source: https://github.com/wrightchin/nanoclaw-playground/tree/main/.claude/skills/add-atomic-chat-tool
Command: npx skills add https://github.com/wrightchin/nanoclaw-playground --skill add-atomic-chat-tool-wrightchin

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @modelcontextprotocol/sdk, zod, and includes scripts (resource) components.

What problem does it solve?

Integrates local Atomic Chat models into NanoClaw by exposing them as MCP tools for containerized agents.

Core Features & Use Cases

  • Exposes a stdio-based MCP server interface to Atomic Chat's /v1 API, enabling listing models and generating chat completions from within the agent.
  • Ships the MCP server source in this skill and wires it into the agent-runner by registering the MCP server, updating the allowlist, and forwarding host configuration.
  • Supports local, offline model usage for secure, private inference without exposing models to external networks.

Quick Start

Install Atomic Chat locally, start the UI, then run the agent so it can discover models and generate responses.

Frequently Asked Questions about add-atomic-chat-tool

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I expose local models as MCP tools for containerized agents?▼

Secure inference with local models keeps AI operations offline and private by forwarding host environment variables to a containerized agent, preventing exposure to external networks while still enabling chat completions and model listing.

How do I integrate local models into a container agent workflow?▼

Integrate local models into a container agent by running a local API on port 1337, wiring the MCP server source into the agent-runner, and forwarding host configuration to enable prompt generation from within the workflow.

Do I need a specific desktop app installed to run local models securely in containers?▼

Yes, you need the Atomic Chat desktop app installed on the host and a running local API at port 1337 to discover models and generate responses securely within the containerized agent.

What code changes are required to register an MCP server in a containerized agent?▼

Registering an MCP server requires updating the agent-runner to register the server, updating the tool allowlist, and forwarding host environment variables to allow the containerized agent to access the local API.

Can I use local models for offline inference without exposing them publicly?▼

Yes, this approach supports local, offline model usage for secure, private inference without exposing models to external networks by keeping the API local and forwarding host configuration to the agent.

What are the limitations of using stdio-based MCP servers for local model integration?▼

Limitations include requiring the desktop app to be actively running on the host for the local API and needing manual code changes to the agent-runner to properly forward host environment variables and register the server.

Related Skills

Explore More Software Engineering Skills→
VincentChong123VincentChong123

add-atomic-chat-tool

Integrate local Atomic Chat models as MCP tools for containerized agents.

Community
Intermediate
BartunekBartunek

add-atomic-chat-tool

Expose local Atomic Chat models as MCP tools for the NanoClaw container agent.

Community
Advanced
involvexinvolvex

typescript-mcp-server-generator

Generate TypeScript MCP server projects with Express or stdio transport and zod validation.

Community
Advanced
ovachieverovachiever
52

serving-llms-vllm

Deploy OpenAI-compatible LLM APIs with vLLM for high-throughput inference.

Community
Advanced