Skip to content

MCP.md

Latest commit

 

History

History
79 lines (56 loc) · 2.13 KB

MCP.md

File metadata and controls

79 lines (56 loc) · 2.13 KB

Model Context Protocol (MCP) Support

The QEM Dashboard supports the Model Context Protocol, allowing AI agents to interact with the dashboard data directly.

Available Tools

  • list_submissions: List active incidents/submissions.
  • get_submission_details: Get details for a specific submission including openQA results.
  • list_blocked: List incidents that are currently blocked.
  • get_repo_status: Get status of various repositories/products.

Example session

$ gemini -y "list all submissions blocked by tests on qem-dashboard"
I will list the incidents that are currently blocked.
The following submissions are currently blocked by tests on qem-dashboard:

1.  **Incident 16860 (perl-Mojolicious)**
    *   **Project:** SUSE:Maintenance:16860
    *   **Blocked by:**
        *   **SLE 12 SP4**: 1 failure

2.  **Incident 29722 (multipath-tools)**
    *   **Project:** SUSE:Maintenance:29722
    *   **Blocked by:**
        *   **SAP-DVD-Updates 15-SP4**: 1 failure in SAP/HA Maintenance
        *   **Server-DVD-Incidents 12-SP6**: 1 failure

Transports

HTTP/SSE (Remote)

The MCP server is available at the /app/mcp endpoint. It uses the standard MCP HTTP/SSE transport.

To use it with an AI agent, configure the agent to point to: https://<dashboard-url>/app/mcp

Note: Authentication via Token might be required depending on the dashboard configuration.

Stdio (Local)

For local interaction or use with desktop agents (like Claude Desktop), a stdio transport script is provided.

Script: script/mcp-stdio

Example configuration for Claude Desktop:

Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "qem-dashboard": {
      "command": "perl",
      "args": ["-Ilib", "script/mcp-stdio"],
      "cwd": "/path/to/qem-dashboard",
      "env": {
        "DASHBOARD_CONF": "dashboard.yml"
      }
    }
  }
}

Development and Testing

To run the MCP server in stdio mode manually for testing:

make run-mcp-stdio

Then you can send JSON-RPC requests, for example:

{"jsonrpc": "2.0", "id": 1, "method": "tools/list"}