Clavis MCP Server

Secure credential management for Claude Desktop and MCP servers.

Features

🔐 Encrypted credential storage (AES-128-CBC + HMAC-SHA256, via Fernet)
🛡️ Server-side credential injection — the raw key never enters the conversation
🔄 Automatic OAuth token refresh
📡 Thin forwarding layer — rate limiting, audit logging, and usage tracking are all handled by the Clavis server

Installation

Terminal

npx @clavisagent/mcp-server

Or install globally:

Terminal

npm install -g @clavisagent/mcp-server

Usage with Claude Desktop

Add the following to your Claude Desktop configuration file (claude_desktop_config.json):

config.json

{
  "mcpServers": {
    "clavis": {
      "command": "npx",
      "args": ["-y", "@clavisagent/mcp-server"],
      "env": {
        "CLAVIS_API_KEY": "eyJ..."
      }
    }
  }
}

Usage with Claude Code

Terminal

claude mcp add clavis -- npx -y @clavisagent/mcp-server

Configuration

Variable	Required	Default	Description
`CLAVIS_API_KEY`	yes	—	Your Clavis JWT, from `POST /v1/auth/login`. Not the `cla_…` key shown at sign-up.
`CLAVIS_API_URL`	no	`https://clavisagent.com`	Base URL of your Clavis instance. Set this for self-hosted deployments.

Available Tools

Tool	Description
`call_service`	Recommended. Make an API call with server-side credential injection — the credential is injected into the upstream request server-side, so the raw key never enters the conversation.
`get_credentials`	Legacy. Returns the raw access token or API key for a named service. Prefer `call_service`.
`list_services`	List all services with stored credentials
`check_credential_status`	Check the status and expiry of credentials for a service

Security note

Prefer call_service over get_credentials. call_service keeps the secret server-side, so a prompt injection has no credential in context to exfiltrate. get_credentials places the raw key in the conversation and exists only for callers that must hold the token themselves.

License

MIT