mcpField
M
mcp-gateway

mcp-gateway

Remote
github.com/civicteamSecurity0Updated 32w ago
Repository

About mcp-gateway

mcp-gateway is Civic's MCP gateway for controlling agent access to tools. It separates identity and permission controls from the agent, with scoped access, audit logs, revocation, and configurable guardrails.

Key features

  1. 01Authenticate agents with Civic tokens
  2. 02Grant access to selected tools
  3. 03Log tool calls and timestamps
  4. 04Revoke access by invalidating tokens
  5. 05Configure rate limits, approvals, and deny lists

Use cases

  1. 01Restrict and audit Gmail sending
  2. 02Constrain Google Calendar actions
  3. 03Limit PostgreSQL queries and redact sensitive columns

MCP support

It provides an MCP endpoint over Streamable HTTP, authenticates with bearer tokens, and applies configured controls to tool access.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://app.civic.com/hub/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

Civic MCP Gateway

Connecting AI Agents to tools and data via the Civic MCP Gateway gives builders access to guardrails, scoped permissions, audit trails, and revocable access when calling MCP tools. Civic separates the permission layer from the AI agent so they can't get around restrictions.

This repository contains integration metadata, configuration files, and client examples. The gateway itself runs at app.civic.com.


What it does

Without an identity layer, any AI agent can call any MCP tool with no record of who authorized it, what it accessed, or how to revoke its access. Civic fixes that:

  • Agent identity — agents authenticate with a Civic-issued bearer token
  • Scoped permissions — grant agents access to specific tools only
  • Audit trail — every tool call is logged with agent identity and timestamp
  • Revocable access — revoke a token and the agent loses access immediately
  • Guardrails — rate limits, approval workflows, and deny lists via the configurator

MCP Endpoint

https://app.civic.com/hub/mcp

Transport: Streamable HTTP Auth: Bearer token (generated at app.civic.com)


Quickstart

1. Get a token

Sign in at app.civic.com and generate a bearer token from the onboarding flow.

2. Configure your MCP client

Claude Desktop — native Streamable HTTP:

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "civic": {
      "url": "https://app.civic.com/hub/mcp"
    }
  }
}

Claude Code — native HTTP transport (no bridge needed):

claude mcp add --transport http civic https://app.civic.com/hub/mcp

Or for stdio via Hub Bridge:

claude mcp add civic "npx @civic/hub-bridge"

Cursor — native Streamable HTTP:

Add to your Cursor MCP config:

{
  "mcpServers": {
    "civic": {
      "url": "https://app.civic.com/hub/mcp"
    }
  }
}

Cline (VS Code) — add to cline_mcp_settings.json:

{
  "mcpServers": {
    "civic": {
      "url": "https://app.civic.com/hub/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_CIVIC_TOKEN"
      }
    }
  }
}

Windsurf — native Streamable HTTP, no bridge required:

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "civic": {
      "serverUrl": "https://app.civic.com/hub/mcp"
    }
  }
}

Windsurf handles OAuth automatically on first connection.

3. Optional: lock to a profile

Append ?profile=your-toolkit to scope the gateway to a specific configuration:

https://app.civic.com/hub/mcp?profile=your-toolkit
Automated agents (token-based auth)

For non-interactive agents that can't complete an OAuth flow, generate a bearer token at app.civic.com and pass it directly:

curl -X POST https://app.civic.com/hub/mcp \
  -H "Authorization: Bearer $CIVIC_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","method":"tools/list","id":1}'

Install from a registry

RegistryLink
Smitherysmithery.ai/servers/@civic/civic-mcp-gateway
ClawHubclawhub.ai/civictechuser/openclaw-civic-skill

Example use cases

Use caseWhat Civic adds
Gmail guardrails & auditBlock outbound sends with secrets, redact PII from email content, log every send
Google Calendar guardrails & auditPrevent event deletion, restrict to business hours, block external attendees
PostgreSQL guardrails & auditBlock destructive queries, enforce read-only access, redact sensitive columns

Documentation

Full documentation at docs.civic.com


Working example

See civicteam/nexus-demo for a full client integration using Next.js, federated auth, and MCP tool visualization.


License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.