Auton — AI skill for Claude Code

auton

An autonomous AI agent you run locally. Give it a task and it works through it — running shell commands, reading and writing files, making HTTP requests, remembering things across sessions, and more. Connect it to Slack, Telegram, or WhatsApp to talk to it from anywhere.

No npm packages required. Just Node.js and an API key.

---

Requirements

• Node.js 18 or later (for native fetch support)
• An OpenAI API key — or any OpenAI-compatible API (Anthropic via proxy, Ollama, Groq, etc.)

Check your Node version:

node --version

---

Setup

Open `config.json` and fill in your API key at minimum:

{
  "llm": {
    "apiKey": "sk-...",
    "apiUrl": "https://api.openai.com/v1",
    "model": "gpt-4o"
  }
}

Everything else in the file has working defaults and can be left alone until you need it.

---

Running the agent

One-shot: give it a single task

node src/agent.js "summarize the files in my Documents folder"

The agent will think through the task, use whatever tools it needs, and stop when it's done. Output prints to the terminal as it works.

Daemon: let it run continuously

node src/agent.js

With no arguments, the agent runs as a persistent background process. Every 30 seconds it checks for scheduled tasks and active goals, then acts on them. Use this when you want it to keep working on something over time or handle scheduled jobs.

---

Talking to it through messaging apps

`src/comms.js` connects the agent to Slack, Telegram, and/or WhatsApp. Configure whichever channels you want in `config.json`, then run:

node src/comms.js

It will start all configured channels at once. If a channel's credentials aren't in the config, it's simply skipped.

---

Slack

Requires **Socket Mode**, which means no public URL needed — the bot connects outbound.

**What you need:**

• A Slack app with Socket Mode enabled
• Bot Token (starts with xoxb-)
• App-Level Token with `conn