Connect AI agents to 85+ tools (Gmail, Google Calendar, PostgreSQL, Slack, and more) through Civic's MCP Gateway with guardrails, scoped permissions, audit trails, and revocable access. Use this skill when you need to call external tools securely with identity, authorization, and logging that the agent can't bypass.
Civic separates the permission layer from the AI agent so agents can't get around restrictions. Every tool call is logged, permissions are user-controlled, and access can be revoked instantly.
{
"mcp": {
"servers": {
"civic": {
"url": "https://app.civic.com/hub/mcp",
"headers": {
"Authorization": "Bearer ${CIVIC_TOKEN}"
}
}
}
}
}
CIVIC_TOKEN (required) — Bearer token generated at app.civic.com during onboardingCIVIC_PROFILE (optional) — Toolkit profile slug appended as ?profile=value to scope the gatewayCIVIC_TOKEN in your environmentClaude Code:
claude mcp add --transport http civic https://app.civic.com/hub/mcp
Claude Desktop / Cursor — add to MCP config:
{
"mcpServers": {
"civic": {
"url": "https://app.civic.com/hub/mcp"
}
}
}
Windsurf — add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"civic": {
"serverUrl": "https://app.civic.com/hub/mcp"
}
}
}
Civic provides access to 85+ MCP servers including:
Civic supports request guardrails (block before execution), response guardrails (redact after execution), parameter guardrails, and transform guardrails — all configurable at app.civic.com.
Examples:
Every tool call is logged with: tool name, input parameters, output, timestamp, and caller identity. Query audit logs at app.civic.com or via the Civic Audit Agent.
Full docs at docs.civic.com