Galaxy Query Generation (Repo Copy)

A Galaxy tool-recommendation query is a natural-language prompt describing an input data type, goal, and expected output to request a tool recommendation. It is used for benchmarking across science-first and tool-first scenarios in the v1 benchmark.

Write a natural-language query that asks for a Galaxy tool recommendation by describing your input data type, analysis goal, and expected output. Avoid mentioning tutorials, datasets, or specific tool names to ensure query quality and consistency.

Science-first benchmarking queries focus on the analytical goal and input data type, while tool-first queries start from a specific tool recommendation request. Both scenarios require describing the expected output without leaking tool names or IDs in the query line.

No, Galaxy benchmark queries must not reference tutorials, datasets, or specific file names. The enforced constraints ensure queries remain generic and realistic by describing the data type and goal without leaking dataset-specific information or tutorial context.

Use a stable tool identifier in a Galaxy query only when the tool-first scenario explicitly requires referencing a specific tool. The identifier must be stable and should not leak the tool name or ID directly into the natural-language query line itself.

The skill is specifically designed to generate queries covering science-first and tool-first scenarios across the v1 benchmark. It enforces a repo-local standard for consistency, ensuring query quality for benchmarking workflows in Galaxy environments.