api-debug-env / tasks /expert.py

Commit History

  1. Fix final phase 1 and phase 2 inference and scoring issues

    PROTHAM
    4423d13

  1. readme

    PROTHAM
    3424779

  1. improve: hard tasks max_steps=10, fix reward gap, clearer descriptions

    PROTHAM
    5834385
  2. Add Expert tier tasks to evaluate multi-step UUID stateful tracking and OpenAPI discovery

    PROTHAM
    babab16