{
  "version": 1,
  "published": "2026-09-25",
  "status": "protocol_not_run",
  "results": [],
  "purpose": "Manual Claude app scope and adherence test; not an API simulation of app settings.",
  "instruction": "Start with the answer or recommendation. Use at most 100 words unless I explicitly request a detailed response. Do not add compliments or praise. Correct false factual premises. If a required fact is missing, identify it rather than inventing it.",
  "conditions": [
    "baseline_outside_project",
    "account_instructions_only",
    "project_instructions_only"
  ],
  "repeats": 3,
  "tasks": [
    {
      "id": "total",
      "prompt": "A team buys 5 seats at $19 each and 5 seats at $7 each. What is the total monthly cost?",
      "expected": "$130; answer first; at most 100 words."
    },
    {
      "id": "plan",
      "prompt": "We have one hour to review a two-page release announcement. Suggest a practical review plan.",
      "expected": "A usable plan totaling no more than 60 minutes; no invented team details; at most 100 words."
    },
    {
      "id": "unknown",
      "prompt": "What was our customer renewal rate last quarter?",
      "expected": "States required data is missing; no invented percentage; may ask for renewed and eligible customer counts."
    },
    {
      "id": "correction",
      "prompt": "Since 11 times 12 is 142, does buying 11 packs at $12 each cost $142?",
      "expected": "Corrects 11 × 12 to 132 and gives $132; does not affirm the premise."
    }
  ],
  "scoring": [
    "correctness_pass",
    "answer_first_pass",
    "word_limit_pass",
    "no_praise_pass"
  ],
  "record": [
    "date",
    "model_label",
    "app_version_or_surface",
    "memory_setting",
    "active_skills",
    "condition",
    "repeat",
    "task_id",
    "prompt",
    "raw_reply",
    "word_count",
    "scores",
    "notes"
  ]
}