MCP Advanced Setup & Reference

Use this page if you need manual MCP wiring or an explicit reference for tools and operational controls.

For the direct local setup flow (envault mcp install) and the delegated envault_agt_ + HITL walkthrough, use Local AI Agents (Claude & Cursor).

Hybrid model: proprietary core, open execution layer

Envault's cloud platform is proprietary. The runtime surfaces that execute agent actions are intentionally open:

• mcp-server/ is fully open source under MIT.
• src/lib/sdk/ is fully open source under MIT.

This gives engineering teams a concrete control surface to verify:

• The HITL interceptor path for mutation requests.
• The delegated short-lived token architecture (envault_agt_ JWT lifecycle).
• The approval-first sequencing that prevents agents from bypassing human approval.

1. Manual Configuration (Standalone MCP Tokens)

Use manual token wiring only when CLI-managed install is not possible (for example isolated DevContainers or custom remote IDE setups).

{
  "mcpServers": {
    "envault": {
      "command": "npx",
      "args": ["-y", "@dinanathdash/envault-mcp-server@latest"],
      "env": {
        "ENVAULT_TOKEN": "<YOUR_ENVAULT_TOKEN>",
        "ENVAULT_BASE_URL": "https://www.envault.tech"
      }
    }
  }
}

{
  "servers": {
    "envault": {
      "command": "npx",
      "args": ["-y", "@dinanathdash/envault-mcp-server@latest"],
      "env": {
        "ENVAULT_TOKEN": "<YOUR_ENVAULT_TOKEN>",
        "ENVAULT_BASE_URL": "https://www.envault.tech"
      }
    }
  }
}

2. HITL Runtime Behavior (Reference)

When an agent requests a mutation, Envault executes the same approval sequence below:

1. Agent Request: The agent uses its MCP envault_push tool to request the mutation.

2. Envault Intercepts: Instead of updating the database, Envault creates a pending_approvals record and returns a 202 Accepted response.

3. Agent Pauses: The agent receives an approval_id and an approval_url. It will pause execution or inform you that an approval is required.

4. Human Verification: You click the approval_url to view the requested changes in the Envault Web Dashboard, or approve it via your terminal:

   envault approve <approval_id>

   Note: The approve command now provides a cryptographic visual diff directly in your terminal, outlining Additions (+), Deletions (-), and Modifications (~). You must be in an interactive terminal to approve the change, otherwise it fails to prevent blind execution.

5. Execution: Once approved, the mutation is securely executed, and the agent proceeds with its task.

The delegated token path is also verifiable in source: agent runtime calls only proceed with scoped, short-lived envault_agt_ credentials, and mutation execution remains gated behind explicit HITL approval.

3. MCP Tools Reference

Once connected, your AI Agent gains access to the following tools:

• Read-Only Tools: envault_status, envault_context, envault_diff, envault_pull
• Mutation Tools (Requires HITL): envault_push, envault_deploy, envault_set_local_key, envault_remove_local_key
• Execution Tools: envault_run (Injects secrets dynamically into a local command)

Security Kill Switches

If an agent behaves unexpectedly or an MCP token is compromised, you do not need to delete the project.

Envault provides instant Kill Switches:

• User Level: Navigate to Settings -> Security and disable "Agent Access". This instantly stops all agents using your tokens across all projects.
• Project Level: Workspace Owners can navigate to Project Settings -> Access Control to disable "Agent Workflows", immediately protecting that specific project from all automated AI modifications.