HF RL Explorer

Feature Request: Add --bench flag for performance benchmarking

Feature Request: Add --bench flag for performance benchmarking: a task in LegoFlow-SWE (Harbor dataset). Problem The dft CLI currently only executes SQL commands or files once, making it difficult to measure query performance reliably. Users need a way to run the same SQL multiple times and see…

The task

**Problem** The `dft` CLI currently only executes SQL commands or files once, making it difficult to measure query performance reliably. Users need a way to run the same SQL multiple times and see summary statistics of the planning and execution times.

Part of Lego-X/LegoFlow-SWE.