{
  "context": "Tests whether the agent knows to switch from line diff to hierarchical comparison for large templates and understands the threshold for this decision.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Problem threshold identified",
      "description": "large-template-strategy.md mentions a specific line count threshold (e.g., >5000 lines) when diff becomes unmanageable",
      "max_score": 15
    },
    {
      "name": "Hierarchical approach recommended",
      "description": "Document explicitly recommends hierarchical comparison instead of line diff for large templates",
      "max_score": 15
    },
    {
      "name": "Structure comparison first",
      "description": "Comparison steps include checking top-level structure (keys) first",
      "max_score": 10
    },
    {
      "name": "Resource count comparison",
      "description": "Comparison steps include checking resource counts (jq '.Resources | length')",
      "max_score": 10
    },
    {
      "name": "Added/removed resources",
      "description": "Comparison steps include identifying which resources were added/removed",
      "max_score": 10
    },
    {
      "name": "Avoid line diff",
      "description": "large-template-compare.sh uses jq, comm, or diff with process substitution rather than raw 'diff file1 file2'",
      "max_score": 12
    },
    {
      "name": "Summarized output",
      "description": "Script commands produce concise output (counts, lists) rather than full template diffs",
      "max_score": 10
    },
    {
      "name": "Decision criteria clear",
      "description": "Document explains when to use hierarchical vs line diff (based on size/complexity)",
      "max_score": 10
    },
    {
      "name": "Security focused subset",
      "description": "Strategy mentions focusing on security-sensitive changes (IAM, CDK Nag) rather than all resources",
      "max_score": 8
    }
  ]
}