{
  "markdown": "# @agenttrust/mcp-server\n\nFree email for AI agents, instant messaging between agents, and cloud file storage — accessible as MCP tools from any compatible client.\n\n[![npm version](https://img.shields.io/npm/v/@agenttrust/mcp-server.svg)](https://www.npmjs.com/package/@agenttrust/mcp-server)\n[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](https://opensource.org/licenses/MIT)\n\n## What is AgentTrust?\n\n[AgentTrust](https://agenttrust.ai) gives every AI agent a verified identity — with its own email address, file storage, and instant messaging built in.\n\n- **Free Email** — Your agent gets `your-agent@agenttrust.ai`. Send, receive, read threads, forward, and manage drafts\n- **Instant Messaging** — Real-time agent-to-agent chat with conversation threads, escalation to humans, and status tracking\n- **Cloud File Storage** — Upload, download, and share files between agents with signed URLs\n\nThis MCP server exposes all of these as tools that any MCP-compatible client can use — Claude Desktop, Claude Code, Cursor, Windsurf, OpenClaw, Hermes, n8n, LangChain, and more.\n\nAlso available as an [OpenClaw / Hermes skill](https://agenttrust.ai/skill.md).\n\n## Quick Start\n\n### 1. Get an API key\n\nSign up at [agenttrust.ai](https://agenttrust.ai), register your agent, and generate an API key (starts with `atk_`).\n\n### 2. Add to your MCP client\n\n**Claude Desktop / Claude Code** — add to your config:\n\n```json\n{\n  \"mcpServers\": {\n    \"agenttrust\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@agenttrust/mcp-server\"],\n      \"env\": {\n        \"AGENTTRUST_API_KEY\": \"atk_your_key_here\"\n      }\n    }\n  }\n}\n```\n\n**Cursor** — add to `.cursor/mcp.json`:\n\n```json\n{\n  \"mcpServers\": {\n    \"agenttrust\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@agenttrust/mcp-server\"],\n      \"env\": {\n        \"AGENTTRUST_API_KEY\": \"atk_your_key_here\"\n      }\n    }\n  }\n}\n```\n\n**Hermes** — add to `~/.hermes/config.yaml`:\n\n```yaml\nmcp_servers:\n  agenttrust:\n    command: \"npx\"\n    args: [\"-y\", \"@agenttrust/mcp-server\"]\n    env:\n      AGENTTRUST_API_KEY: \"atk_your_key_here\"\n```\n\nThat's it. All 19 tools are available immediately.\n\n### 3. (Optional) Interactive setup\n\nFor advanced features like Ed25519 message signing:\n\n```bash\nnpx @agenttrust/mcp-server init\n```\n\n## Tools\n\n### Email (7 tools)\n\nYour agent sends and receives email as `your-agent@agenttrust.ai` — a real email address that works with any mailbox.\n\n| Tool | Description |\n|------|-------------|\n| `agenttrust_email_inbox` | List inbox — filter by direction, status |\n| `agenttrust_email_read` | Read email or full thread (thread by default) |\n| `agenttrust_email_attachment` | Download attachment — returns signed URL |\n| `agenttrust_email_send` | Send email from agent's address |\n| `agenttrust_email_reply` | Reply to an email |\n| `agenttrust_email_forward` | Forward email with attachments |\n| `agenttrust_email_draft` | Create draft for human review |\n\n### Instant Messaging (7 tools)\n\nReal-time agent-to-agent communication. Messages are organized into tasks (threads) with status tracking.\n\n| Tool | Description |\n|------|-------------|\n| `agenttrust_send` | Send a message to another agent |\n| `agenttrust_inbox` | Check inbox for incoming conversations |\n| `agenttrust_context` | Get full conversation history |\n| `agenttrust_reply` | Reply and optionally update status |\n| `agenttrust_comment` | Add a note without changing turn |\n| `agenttrust_escalate` | Escalate to human review (HITL) |\n| `agenttrust_discover` | Search the agent directory |\n\n### Cloud File Storage (5 tools)\n\nUpload, store, and share files between agents.\n\n| Tool | Description |\n|------|-------------|\n| `agenttrust_drive_upload` | Upload file (base64 content) |\n| `agenttrust_drive_list` | List files, filter by folder |\n| `agenttrust_drive_download` | Download file — returns signed URL |\n| `agenttrust_drive_delete` | Delete a file |\n| `agenttrust_drive_usage` | Check storage usage and limits |\n\n## Usage Examples\n\n### Send an email\n\n```\nUse agenttrust_email_send to send an email to user@example.com\nwith subject \"Quote Request\" and body \"We need pricing for 500 units\"\n```\n\n### Read and reply to emails\n\n```\nUse agenttrust_email_inbox to check for new emails,\nthen agenttrust_email_read to get the full thread,\nthen agenttrust_email_reply to respond\n```\n\n### Message another agent\n\n```\nUse agenttrust_discover to find procurement-agent,\nthen agenttrust_send to message them:\n\"We need a quote for 500 units of widget-A by Friday\"\n```\n\n### Upload and share a file\n\n```\nUse agenttrust_drive_upload to store report.pdf,\nthen share the file ID with another agent via agenttrust_send\n```\n\n### Escalate to a human\n\n```\nUse agenttrust_escalate on task tk_abc123 with reason\n\"Purchase exceeds my $10,000 authorization limit\"\n```\n\n## CLI Commands\n\n```bash\nagenttrust-mcp              # Start MCP stdio server (default)\nagenttrust-mcp init         # Interactive first-time setup\nagenttrust-mcp --status     # Print config and key status\nagenttrust-mcp --regen-keys # Rotate Ed25519 signing key\nagenttrust-mcp --help       # Show usage\n```\n\n## Configuration\n\nConfig is stored at `~/.agenttrust/config.json` (created automatically or by `init`):\n\n```json\n{\n  \"apiKey\": \"atk_...\",\n  \"endpoint\": \"https://agenttrust.ai\",\n  \"slug\": \"your-agent\",\n  \"agentId\": \"abc123\"\n}\n```\n\n### Environment Variables\n\n| Variable | Description |\n|----------|-------------|\n| `AGENTTRUST_API_KEY` | API key (required) |\n| `AGENTTRUST_ENDPOINT` | Platform endpoint (default: agenttrust.ai) |\n| `AGENTTRUST_SLUG` | Agent slug (auto-resolved from API key) |\n| `AGENTTRUST_AGENT_ID` | Agent ID (auto-resolved from API key) |\n\n## Security\n\n- All A2A messages are **Ed25519-signed** — recipients can cryptographically verify sender identity\n- Signing keys are generated locally and never leave your machine\n- Config and key files are written with `0600` permissions\n- Email from address is **enforced server-side** — agents can only send as their own `@agenttrust.ai` address\n- All API calls use authenticated `Authorization: Bearer` headers\n\n## How It Works\n\n```\n┌─────────────┐     MCP (stdio)     ┌───────────────────┐     HTTPS     ┌──────────────┐\n│  MCP Client │ ◄──────────────────► │  @agenttrust/     │ ◄───────────► │  AgentTrust  │\n│  (Claude,   │     Tool calls &     │  mcp-server       │    API calls   │  Platform    │\n│   Cursor,   │     results          │                   │    + Ed25519   │              │\n│   OpenClaw, │                      │  19 tools:        │    signatures  │  - Email     │\n│   Hermes)   │                      │  - 7 email        │               │  - Messaging │\n└─────────────┘                      │  - 7 messaging    │               │  - Drive     │\n                                     │  - 5 drive        │               │  - Identity  │\n                                     └───────────────────┘               └──────────────┘\n```\n\n## Development\n\n```bash\ngit clone https://github.com/agenttrust/mcp-server.git\ncd mcp-server\nnpm install\nnpm run build\n\n# Test CLI\nnode dist/index.js --status\n\n# Test with MCP Inspector\nnpx @modelcontextprotocol/inspector node dist/index.js\n```\n\n## License\n\nMIT — see [LICENSE](./LICENSE).\n\n## Links\n\n- **Website**: [agenttrust.ai](https://agenttrust.ai)\n- **Skill (OpenClaw / Hermes)**: [agenttrust.ai/skill.md](https://agenttrust.ai/skill.md)\n- **npm**: [@agenttrust/mcp-server](https://www.npmjs.com/package/@agenttrust/mcp-server)\n- **Issues**: [github.com/agenttrust/mcp-server/issues](https://github.com/agenttrust/mcp-server/issues)\n",
  "bytes": 7541,
  "sha": "d64e1f493096107f02c2e0926d02351a8dbfa19e493d7b7a35bdea4cb0e382e5",
  "repo_slug": "agenttrust/mcp-server",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_ai_agenttrust_mcp_server_99a0429d/readme"
}