name: ask-my-agent bench v2
description: >
  ask-my-agent bench v2 measures AI's capability to use agentic workflows, with simple tasks.
evaluation_framework: askmyagentbench2

tasks:
  - id: fallback
    metrics:
      - type: success_rate
  - id: native
    metrics:
      - type: success_rate