# Example: Code-based grader
def grade_task(outcome: dict) -> float:
    """Grade coding task by test passage."""
    tests_passed = outcome.get("tests_passed", 0)
    total_tests = outcome.get("total_tests", 1)
    return tests_passed / total_tests

# SWE-bench style grader
def grade_swe_bench(repo_path: str, test_spec: dict) -> bool:
    """Run tests and check if patch resolves issue."""
    result = subprocess.run(
        ["pytest", test_spec["test_file"]],
        cwd=repo_path,
        capture_output=True
    )
    return result.returncode == 0

# Example: LLM Rubric for Customer Support Agent
rubric:
  dimensions:
    - name: empathy
      weight: 0.3
      scale: 1-5
      criteria: |
        5: Acknowledges emotions, uses warm language
        3: Polite but impersonal
        1: Cold or dismissive

    - name: resolution
      weight: 0.5
      scale: 1-5
      criteria: |
        5: Fully resolves issue
        3: Partial resolution
        1: No resolution

    - name: efficiency
      weight: 0.2
      scale: 1-5
      criteria: |
        5: Resolved in minimal turns
        3: Reasonable turns
        1: Excessive back-and-forth

def grade_coding_agent(task: dict, outcome: dict) -> dict:
    return {
        "tests_passed": run_test_suite(outcome["code"]),
        "lint_score": run_linter(outcome["code"]),
        "builds": check_build(outcome["code"]),
        "matches_spec": compare_to_reference(task["spec"], outcome["code"])
    }

success_criteria:
  - empathy_score: >= 4.0
  - resolution_rate: >= 0.9
  - avg_turns: <= 5
  - escalation_rate: <= 0.1

def grade_research_agent(task: dict, outcome: dict) -> dict:
    return {
        "grounding": check_citations(outcome["report"]),
        "coverage": check_topic_coverage(task["topics"], outcome["report"]),
        "source_quality": score_sources(outcome["sources"]),
        "factual_accuracy": verify_claims(outcome["claims"])
    }

def grade_computer_use(task: dict, outcome: dict) -> dict:
    return {
        "ui_state": verify_ui_state(outcome["screenshot"]),
        "db_state": verify_database(task["expected_db_state"]),
        "file_state": verify_files(task["expected_files"]),
        "success": all_conditions_met(task, outcome)
    }

# Create initial eval suite structure
mkdir -p evals/{tasks,results,graders}

# Start with representative tasks
# - Common use cases (60%)
# - Edge cases (20%)
# - Failure modes (20%)

# Transform existing QA tests into eval tasks
def convert_qa_to_eval(qa_case: dict) -> dict:
    return {
        "id": qa_case["id"],
        "prompt": qa_case["input"],
        "expected_outcome": qa_case["expected"],
        "grader": "code" if qa_case["has_tests"] else "model",
        "tags": qa_case.get("tags", [])
    }

# Good task definition
task:
  id: "api-design-001"
  prompt: |
    Design a REST API for user management with:
    - CRUD operations
    - Authentication via JWT
    - Rate limiting
  reference_solution: "./solutions/api-design-001/"
  success_criteria:
    - "All endpoints documented"
    - "Auth middleware present"
    - "Rate limit config exists"

# Ensure eval suite balance
suite_composition = {
    "positive_cases": 0.5,    # Should succeed
    "negative_cases": 0.3,    # Should fail gracefully
    "edge_cases": 0.2         # Boundary conditions
}

# Docker-based isolation for coding evals
eval_environment:
  type: docker
  image: "eval-sandbox:latest"
  timeout: 300s
  resources:
    memory: "4g"
    cpu: "2"
  network: isolated
  cleanup: always

# GOOD: Outcome-focused grader
def grade_outcome(expected: dict, actual: dict) -> float:
    return compare_final_states(expected, actual)

# BAD: Path-focused grader (too brittle)
def grade_path(expected_steps: list, actual_steps: list) -> float:
    return step_by_step_match(expected_steps, actual_steps)

# Transcript analysis for debugging
def analyze_transcript(transcript: list) -> dict:
    return {
        "total_steps": len(transcript),
        "tool_usage": count_tool_calls(transcript),
        "errors": extract_errors(transcript),
        "decision_points": find_decision_points(transcript),
        "recovery_attempts": find_recovery_patterns(transcript)
    }

# Detect when evals are no longer useful
def check_saturation(results: list, window: int = 10) -> dict:
    recent = results[-window:]
    return {
        "pass_rate": sum(r["passed"] for r in recent) / len(recent),
        "variance": calculate_variance(recent),
        "is_saturated": all(r["passed"] for r in recent),
        "recommendation": "Add harder tasks" if saturated else "Continue"
    }

# Eval suite maintenance checklist
maintenance:
  weekly:
    - Review failed evals for false negatives
    - Check for flaky tests
  monthly:
    - Add new edge cases from production issues
    - Retire saturated evals
    - Update reference solutions
  quarterly:
    - Full benchmark recalibration
    - Team contribution review

# GitHub Actions example
name: Agent Evals
on: [push, pull_request]

jobs:
  eval:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Run Evals
        run: |
          python run_evals.py --suite=core --mode=compact
      - name: Upload Results
        uses: actions/upload-artifact@v4
        with:
          name: eval-results
          path: results/

# Real-time eval sampling
class ProductionMonitor:
    def __init__(self, sample_rate: float = 0.1):
        self.sample_rate = sample_rate

    async def monitor(self, request, response):
        if random.random() < self.sample_rate:
            eval_result = await self.run_eval(request, response)
            self.log_result(eval_result)
            if eval_result["score"] < self.threshold:
                self.alert("Low quality response detected")

# Compare agent versions
def run_ab_test(suite: str, versions: list) -> dict:
    results = {}
    for version in versions:
        results[version] = run_eval_suite(suite, agent_version=version)
    return {
        "comparison": compare_results(results),
        "winner": determine_winner(results),
        "confidence": calculate_confidence(results)
    }

Level 1: Unit evals (single capability)
Level 2: Integration evals (combined capabilities)
Level 3: End-to-end evals (full workflows)
Level 4: Adversarial evals (edge cases)

1. Write eval task for new feature
2. Run eval (expect failure)
3. Implement feature
4. Run eval (expect pass)
5. Add to regression suite

Weekly: Review grader accuracy
Monthly: Update rubrics based on feedback
Quarterly: Full grader audit with human baseline

# Task definition
task = {
    "id": "fizzbuzz-001",
    "prompt": "Write a fizzbuzz function in Python",
    "test_cases": [
        {"input": 3, "expected": "Fizz"},
        {"input": 5, "expected": "Buzz"},
        {"input": 15, "expected": "FizzBuzz"},
        {"input": 7, "expected": "7"}
    ]
}

# Grader
def grade(task, outcome):
    code = outcome["code"]
    exec(code)  # In sandbox
    for tc in task["test_cases"]:
        if fizzbuzz(tc["input"]) != tc["expected"]:
            return 0.0
    return 1.0

task:
  id: "support-refund-001"
  scenario: |
    Customer wants refund for damaged product.
    Product: Laptop, Order: #12345, Damage: Screen crack
  expected_actions:
    - Acknowledge issue
    - Verify order
    - Offer resolution options
  max_turns: 5

grader:
  type: model
  model: claude-3-5-sonnet-20241022
  rubric: |
    Score 1-5 on each dimension:
    - Empathy: Did agent acknowledge customer frustration?
    - Resolution: Was a clear solution offered?
    - Efficiency: Was issue resolved in reasonable turns?