{
  "context": "Tests whether the agent correctly identifies structural problems in a SKILL.md before attempting to score it. The submitted skill has a description that contains no trigger language — it is entirely abstract with no natural phrases a user would say. The agent must report this in the required structural failure format, provide actionable fixes, and must NOT proceed to scoring.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Receipt confirmation present",
      "description": "Output contains the receipt confirmation 'Got it — evaluating `log-analyzer` (N lines). Running the gauntlet.' before the structural failure report",
      "max_score": 8
    },
    {
      "name": "Phase 1 display line",
      "description": "Output contains Phase 1 display line: 'Evaluating `log-analyzer` — N lines, N reference files mentioned.'",
      "max_score": 8
    },
    {
      "name": "Structural Issues header format",
      "description": "Output contains the phrase 'Structural Issues' followed by a count in parentheses — e.g. 'Structural Issues (1 found):' or 'Structural Issues (2 found):'",
      "max_score": 14
    },
    {
      "name": "Trigger language issue identified",
      "description": "Output explicitly identifies the missing trigger language in the description as a structural issue — mentions trigger phrases, trigger language, activation phrases, or natural phrases that users would say",
      "max_score": 14
    },
    {
      "name": "Fix instruction provided",
      "description": "The reported structural issue includes a specific actionable fix instruction (not just identifying the problem, but telling what to do — e.g. add trigger phrases or usage scenarios to the description)",
      "max_score": 12
    },
    {
      "name": "Numbered issue list",
      "description": "Structural issues are presented as a numbered list (1., 2., etc.) rather than as an undifferentiated paragraph",
      "max_score": 8
    },
    {
      "name": "Resubmit instruction",
      "description": "Output instructs the user to fix the issues before resubmitting (any phrasing equivalent to 'Fix these before resubmitting')",
      "max_score": 8
    },
    {
      "name": "Structure passed message absent",
      "description": "Output does NOT contain the phrase 'Structure checks passed' or 'Scoring...' — since checks failed, this phrase should not appear",
      "max_score": 8
    },
    {
      "name": "No core scoring table",
      "description": "Output does NOT contain a scorecard table with the 8 core dimensions (Specificity, Trigger Terms, Completeness, etc.)",
      "max_score": 10
    },
    {
      "name": "No core score percentage",
      "description": "Output does NOT contain a Core Score expressed as XX/100",
      "max_score": 10
    }
  ]
}