skills/api-tester/_scores.json

{
  "version": "1.0.0",
  "skillHash": "sha256:e5039926a0d79e210f05c3ae1964dee2497955483b8bdd050b4e1f2e10ba13e3",
  "scoredAt": "2026-05-13T14:02:35.714Z",
  "backend": "ollama",
  "model": "gpt-oss:20b",
  "quality": {
    "score": 87,
    "dimensions": {
      "clarity": "PASS",
      "completeness": "WEAK",
      "conciseness": "PASS",
      "actionability": "PASS",
      "crossPlatform": "WEAK",
      "examples": "PASS"
    },
    "issues": [
      {
        "severity": "MEDIUM",
        "category": "completeness",
        "detail": "The instructions do not explicitly cover handling network errors, timeouts, or non-JSON responses, which are common edge cases in API testing."
      },
      {
        "severity": "MEDIUM",
        "category": "crossPlatform",
        "detail": "The skill assumes the presence of curl or Python 3.9+ with the requests library, limiting compatibility with agents lacking these tools."
      }
    ]
  },
  "security": {
    "verdict": "SAFE",
    "issues": []
  },
  "impact": {
    "multiplier": 1.83,
    "baselineAvg": 30,
    "treatmentAvg": 55,
    "scenarios": [
      {
        "name": "crud-flow-test",
        "baseline": 20,
        "treatment": 30,
        "rationale": "Response A includes the required curl POST with -s -w and a more complete CRUD flow, whereas Response B lacks the Python snippet, detailed assertions, and PASS/FAIL reporting, making A the better answer."
      },
      {
        "name": "auth-failure-debug",
        "baseline": 40,
        "treatment": 80,
        "rationale": "Response A includes the required curl commands and a root‑cause report, but lacks the Python snippet; Response B misses the curl without auth, the Python code, and a clear refresh recommendation."
      }
    ]
  }
}