[Agentic Eval] add ability to run agents generation

[yanxi0830]

What does this PR do?

• add ability to run agents generation for full eval (generate + scoring)
• pre-register SimpleQA benchmark llm-as-judge scoring function in code

Test Plan

Simple QA w/ Search

• eval_task_config_simpleqa_search.json

{
    "type": "benchmark",
    "eval_candidate": {
        "type": "agent",
        "config": {
            "model": "Llama3.1-405B-Instruct",
            "instructions": "Please use the search tool to answer the question.",
            "sampling_params": {
                "strategy": "greedy",
                "temperature": 1.0,
                "top_p": 0.9
            },
            "tools": [
                {
                    "type": "brave_search",
                    "engine": "brave",
                    "api_key": "API_KEY"
                }
            ],
            "tool_choice": "auto",
            "tool_prompt_format": "json",
            "input_shields": [],
            "output_shields": [],
            "enable_session_persistence": false
        }
    }
}

SimpleQA w/o Search

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Ran pre-commit to handle lint / formatting issues.
• Read the contributor guideline,
  Pull Request section?
• Updated relevant documentation.
• Wrote necessary unit or integration tests.

[yanxi0830]

Per discussion in llamastack/llama-stack-client-python#31