Files
gstack/evals/parity/contracts/benchmark-models.json
T

29 lines
1.6 KiB
JSON

{
"source": "benchmark-models",
"tree": "qa",
"public_mode": "Report",
"mode": "model-benchmark",
"mandatory": false,
"visibility": "internal",
"replacement": "$qa --mode Report --module benchmark-models",
"source_path": "benchmark-models/SKILL.md.tmpl",
"base_sha": "bb57306d98c97011b0919c6132705a15b1579781",
"blob_sha": "034cda182406dc04a82c4336ac3ebc36b5fc41b1",
"normalized_render_sha256": "2ef0679d45f21bacc09cd774ff96bb3b82853a8e89d8606847c4f9416a47a48b",
"target": "skills/qa/references/legacy/benchmark-models.md",
"overlays": [
679
],
"contract": {
"question_order": "Preserve the source workflow order; gather prerequisites before consequential questions.",
"pressure": "Preserve the source forcing questions, recommendation pressure, and one-question-at-a-time cadence.",
"smart_skips": "Skip only when the source condition is false, and name every skipped module with evidence.",
"stop_approval_gates": "Preserve every STOP, hard gate, approval boundary, and no-mutation-before-approval rule.",
"evidence": "Ground conclusions in inspected code, commands, browser/device observations, or source artifacts.",
"artifacts": "Produce every report, plan, log, screenshot, manifest, or handoff required by the source.",
"mutation": "Use the source mutation boundary; never broaden writes, commits, pushes, merges, or deploys.",
"exit": "Preserve source completion checks, unresolved-decision reporting, and explicit blocked exits.",
"voice": "Direct builder voice; match the user language and retain source-specific tone constraints."
}
}