skills/aws-sqs/_scores.json

{
  "version": "1.0.0",
  "skillHash": "sha256:f9c847c3baa27982934e7724f467235a868b36f421acc73bfe95d1b9d180ac4e",
  "scoredAt": "2026-05-13T14:29:29.438Z",
  "backend": "ollama",
  "model": "gpt-oss:20b",
  "quality": {
    "score": 77,
    "dimensions": {
      "clarity": "PASS",
      "completeness": "WEAK",
      "conciseness": "PASS",
      "actionability": "PASS",
      "crossPlatform": "FAIL",
      "examples": "PASS"
    },
    "issues": [
      {
        "severity": "MEDIUM",
        "category": "completeness",
        "detail": "The skill does not cover error handling or edge cases for queue operations."
      },
      {
        "severity": "HIGH",
        "category": "crossPlatform",
        "detail": "The skill is vendor-locked to AWS and requires AWS CLI and boto3, limiting cross-platform applicability."
      }
    ]
  },
  "security": {
    "verdict": "SAFE",
    "issues": []
  },
  "impact": {
    "multiplier": 2.43,
    "baselineAvg": 35,
    "treatmentAvg": 85,
    "scenarios": [
      {
        "name": "setup-fifo-queue-with-dlq-and-batch-send",
        "baseline": 20,
        "treatment": 70,
        "rationale": "Response B meets most rubric items except ContentBasedDeduplication, while Response A fails on queue names, attributes, and redrive policy."
      },
      {
        "name": "batch-receive-process-delete-monitor",
        "baseline": 50,
        "treatment": 100,
        "rationale": "Response B meets all rubric requirements, while Response A uses short polling and single deletes, missing key elements."
      }
    ]
  }
}