docs
evals
scenario-1
scenario-10
scenario-2
scenario-3
scenario-4
scenario-5
scenario-6
scenario-7
scenario-8
scenario-9
{
"context": "Evaluates how well the solution surfaces and formats the punycode library's version by relying on the package's own metadata rather than hard-coded values. Scoring emphasizes using punycode as the single source of truth for the reported version string.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Source is punycode.version",
"description": "Version retrieval reads directly from punycode.version (or the identical exported version field) without hard-coding literals or manually parsing package.json.",
"max_score": 40
},
{
"name": "Exact value",
"description": "The exposed getter returns the exact string provided by punycode.version with no modification, trimming, or fallback defaults.",
"max_score": 20
},
{
"name": "Formatting uses source",
"description": "Formatted outputs (e.g., label + version) incorporate the same punycode.version value rather than duplicating or re-deriving a separate version string.",
"max_score": 20
},
{
"name": "Semver fidelity",
"description": "Returned version text preserves the semver structure emitted by punycode.version (major.minor.patch) without reformatting or token loss.",
"max_score": 10
},
{
"name": "Consistent reads",
"description": "Repeated calls still source the value from punycode.version so updates to the dependency reflect immediately; no stale caches of another value.",
"max_score": 10
}
]
}