AI Agent Integration

Integrate Talos and Omni documentation with AI agents using llms.txt, skill.md, or MCP.

SideroLabs publishes machine-readable documentation metadata that allows AI agents and developer tools to understand Talos Linux and Omni capabilities programmatically.

These integration points can be used by agent frameworks, MCP-compatible tools, IDE extensions, or custom automation.

Available integration endpoints

Depending on your tooling and integration model, you can use one or more of the following endpoints.

The llms.txt file

The llms.txt file is a structured documentation index.

It lists all the available documentation pages so agents can discover relevant content before answering questions.

You can see the llms.txt file at:

https://docs.siderolabs.com/llms.txt

The skill.md file

The skill.md file is a capability definition that describes supported operations, workflows, architectural constraints, and expected inputs and outputs.

Agents can use this file to understand:

• Supported actions
• Required configuration inputs
• Workflow sequencing
• Operational constraints

You can download the skill.md file in your personal Claude skills directory:

mkdir -p ~/.claude/skills/siderolabs
curl -L https://docs.siderolabs.com/skill.md \
  -o ~/.claude/skills/siderolabs/SKILL.md

Restart Claude Code after downloading the file.

Alternatively, if your agent supports the Skills CLI, you can install the skill using:

npx skills add https://docs.siderolabs.com

This command loads the SideroLabs capability definition into the agent’s context.

MCP endpoint

SideroLabs exposes its documentation through a remote Model Context Protocol (MCP) server.

This enables MCP-compatible clients to query the documentation dynamically using the standard MCP interface.

To connect, register the following URL as a remote MCP server in your client configuration:

https://docs.siderolabs.com/mcp

For example:

To register the SideroLabs documentation as an MCP server in OpenCode or Claude Code:

Claude Code

The recommended way to add the server is with the claude mcp add command. Choose the scope that best fits how you want to use it:

# Available in all your projects (private to you):
claude mcp add --scope user --transport http siderolabs-docs https://docs.siderolabs.com/mcp

# Shared with your team (writes .mcp.json in the repository root):
claude mcp add --scope project --transport http siderolabs-docs https://docs.siderolabs.com/mcp

After adding the server, verify the connection:

claude mcp list

If the server is configured correctly, it appears as ✔ Connected.

When you use the --scope project option, Claude Code creates a .mcp.json file in the root of your repository with the following configuration:

{
  "mcpServers": {
    "siderolabs-docs": {
      "type": "http",
      "url": "https://docs.siderolabs.com/mcp"
    }
  }
}

OpenCode

Add the MCP server configuration to your OpenCode config file (opencode.json or opencode.jsonc):

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "siderolabs-docs": {
      "type": "remote",
      "url": "https://docs.siderolabs.com/mcp",
      "enabled": true
    }
  }
}

Once added, the SideroLabs documentation tools are available to the agent alongside its built-in tools.