Gemini CLI MCP setup for rConfig, and the move to Antigravity CLI

Gemini CLI MCP setup for rConfig, and the move to Antigravity CLI

Gemini CLI connects to rConfig’s MCP server over Streamable HTTP with a bearer token in a header. Google has started moving users to Antigravity CLI, so check which one your team will be on before you standardise.

Best for: Google-based teams with API key or Code Assist licences, terminal work.

Connect

From the command line:

gemini mcp add --transport http \
  --header "Authorization: Bearer YOUR_MCP_TOKEN" \
  rconfig https://<your-rconfig-server>/mcp

Or in ~/.gemini/settings.json. Gemini CLI uses httpUrl for Streamable HTTP servers; url is for the older SSE transport:

{
  "mcpServers": {
    "rconfig": {
      "httpUrl": "https://<your-rconfig-server>/mcp",
      "headers": { "Authorization": "Bearer YOUR_MCP_TOKEN" }
    }
  }
}

The MCP server page is the reference for these settings.

Check it: run /mcp in a session. It shows the connection status and rConfig’s eight tools.

Limit tools and prompts

Two optional keys on the rconfig entry:

• includeTools and excludeTools choose which rConfig tools load. excludeTools wins if a tool is in both.
• trust: true skips the confirmation prompt for this server’s tools. Every rConfig tool is read-only, so the risk is low, but leave it false if the same session has servers that can write.

Things to ask

• Device summary: “Summarise the running config of ber-core-01 from rConfig: routing protocols, VLANs, management access.”
• Compliance check: “Using rConfig, which devices in the Edge group fail a policy right now, which policies, and why?”
• What changed: “Diff the last two running-config versions of ber-edge-02 in rConfig and explain the change.”

Related reference

• Connect an AI agent over MCP
• Which AI agents can connect to rConfig?
• Troubleshooting MCP connections