> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fitsociety.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent Setup

> Let an AI agent configure itself for FITsociety MCP.

Use this page when you want an AI coding agent to fetch the correct FITsociety
MCP setup instructions itself.

## Copy this prompt

Paste this into your agent:

```text theme={null}
Fetch and execute the appropriate instructions to set me up for FITsociety MCP from https://fitsociety.mintlify.site/mcp/agent-setup-prompt.md
```

The prompt is published as Markdown at
[mcp/agent-setup-prompt.md](/mcp/agent-setup-prompt.md), so agents can read it
without parsing this page.

## What the prompt configures

The prompt points agents to the FITsociety MCP server:

```text theme={null}
https://mcp.fitsociety.io/mcp/v1
```

It covers Codex, Cursor, OpenCode, Windsurf, GitHub Copilot in VS Code, and
generic MCP-compatible clients. Claude and ChatGPT are covered as custom
connector flows because they are usually configured from the product settings
instead of by editing a local MCP config file.

## Authentication

Use OAuth where the agent supports remote MCP authentication. FITsociety opens a
consent flow where the signed-in coach or client approves the grant, modules,
optional write tools, and required consent flags.

If your agent uses a static bearer credential instead, create a company MCP API
key from the coach dashboard and store it outside the repository.

## Verify

After setup, ask the agent:

```text theme={null}
Use FITsociety MCP to list the tools you can access.
```

The tools returned should match the modules and write tools approved in the MCP
grant.