MCP server gains query tools and read-only mode

Developers using the Trigger.dev MCP server can now give their AI assistants significantly more power to analyze runs and metrics. The documentation has been updated to include a suite of new tools that allow assistants to run TRQL queries, inspect available tables, and fetch data from built-in dashboards.

With these new capabilities, users can ask their AI questions like "How many runs failed in the last 7 days?" or "What are the most expensive tasks?" and the assistant will leverage the query schema to fetch the correct data. To balance this new power, a read-only mode has been documented, allowing users to restrict their AI from deploying or triggering tasks, ensuring it acts strictly as an analytical tool.

This update, depending on ongoing backend work, also brings documentation for profile management and dev server controls directly into the MCP ecosystem.