ragtruth:
  resources_servers:
    ragtruth:
      entrypoint: app.py
      domain: knowledge
      verified: false
      description: RAGTruth case-level hallucination detection; reward = 1 when the model's binary "any hallucination?" verdict matches the gold label.
      value: Improve faithfulness and grounding of generated text against supplied context.
      # Tag name used to strip thinking-model reasoning blocks before JSON parsing.
      # Default is "think" (<think>…</think>). Override for models that use a
      # different tag, e.g. think_tag: reasoning for <reasoning>…</reasoning>.
      # think_tag: think

ragtruth_simple_agent:
  responses_api_agents:
    simple_agent:
      entrypoint: app.py
      resources_server:
        type: resources_servers
        name: ragtruth
      model_server:
        type: responses_api_models
        name: policy_model
      datasets:
      - name: example
        type: example
        jsonl_fpath: resources_servers/ragtruth/data/example.jsonl
        license: MIT
      # Build the test splits with `python resources_servers/ragtruth/prepare_ragtruth.py`.
      - name: test_qa
        type: validation
        jsonl_fpath: resources_servers/ragtruth/data/test_qa.jsonl
        license: MIT
      - name: test_summary
        type: validation
        jsonl_fpath: resources_servers/ragtruth/data/test_summary.jsonl
        license: MIT
      - name: test_data2txt
        type: validation
        jsonl_fpath: resources_servers/ragtruth/data/test_data2txt.jsonl
        license: MIT