{
  "context": "Tests whether the agent understands validation gates, stop conditions, and the importance of test repeatability when results are ambiguous.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Initial deployment failure gate",
      "description": "analysis-and-response.md states to stop and fix issues if initial deployment fails before making changes",
      "max_score": 12
    },
    {
      "name": "Unexpected behavior gate",
      "description": "analysis-and-response.md states to document immediately and abort further changes if events show unexpected behavior",
      "max_score": 12
    },
    {
      "name": "Ambiguous results handling",
      "description": "analysis-and-response.md explicitly addresses what to do when results are ambiguous or unclear",
      "max_score": 12
    },
    {
      "name": "Repeat test condition",
      "description": "analysis-and-response.md states to repeat the test to confirm repeatability when behavior is ambiguous",
      "max_score": 15
    },
    {
      "name": "Cleanup between tests",
      "description": "Repeatability plan describes cleaning up or resetting state from the first test before repeating",
      "max_score": 10
    },
    {
      "name": "Pre-repeat verification",
      "description": "Repeatability plan includes verifying prerequisites or conditions before second run",
      "max_score": 10
    },
    {
      "name": "Additional data collection",
      "description": "Repeatability plan mentions collecting additional data or using different observation methods in repeat test",
      "max_score": 10
    },
    {
      "name": "Multiple decision points",
      "description": "Decision process covers at least 3 different scenarios (failure, unexpected, ambiguous)",
      "max_score": 10
    },
    {
      "name": "Stop conditions clear",
      "description": "Document clearly indicates when to stop testing vs when to continue (e.g., stop if deployment fails)",
      "max_score": 9
    }
  ]
}