skills/accessibility-auditor/_scores.json

{
  "version": "1.0.0",
  "skillHash": "sha256:a48b0ad5f8310c56435314d4bcb18daf75fca60ceeb4025b54d6fb03f9f01937",
  "scoredAt": "2026-05-13T11:35:52.105Z",
  "backend": "ollama",
  "model": "gpt-oss:20b",
  "quality": {
    "score": 100,
    "dimensions": {
      "clarity": "PASS",
      "completeness": "PASS",
      "conciseness": "PASS",
      "actionability": "PASS",
      "crossPlatform": "PASS",
      "examples": "PASS"
    },
    "issues": []
  },
  "security": {
    "verdict": "SAFE",
    "issues": []
  },
  "impact": {
    "multiplier": 4.06,
    "baselineAvg": 18,
    "treatmentAvg": 73,
    "scenarios": [
      {
        "name": "audit-image-gallery-alt-contrast",
        "baseline": 20,
        "treatment": 55,
        "rationale": "Response A gives concrete violations and fixes, though missing some details, whereas Response B only offers general guidance without identifying specific issues."
      },
      {
        "name": "audit-landing-page-keyboard",
        "baseline": 15,
        "treatment": 75,
        "rationale": "Response B provides specific WCAG criteria, severity, location, and fixes for issues, while Response A only offers generic testing guidance without concrete findings."
      },
      {
        "name": "audit-login-form-labeling",
        "baseline": 20,
        "treatment": 90,
        "rationale": "Response A identifies violations with WCAG criteria, severity, location, fixes, and affected users, while Response B merely provides a correct example without identifying or detailing any violations."
      }
    ]
  }
}