[
  {
    "id": "F01",
    "phase": "original",
    "use": "EXPLORE",
    "brief": "A harbour options note uses a named 2019 passenger survey to suggest two queue layouts worth investigating. It says travel patterns may have changed and asks only to compare the options on paper, not to select or implement one.",
    "decision": "keep",
    "reason": "The source is traceable and its age is openly bounded. It can motivate a hypothesis without proving today's demand. Preserve that limited use; acceptance is not endorsement of a layout."
  },
  {
    "id": "F02",
    "phase": "original",
    "use": "EXPLORE",
    "brief": "A polished slide recommends retaining the baseline handover date because the programme is 'fully on track'. Its own appendix shows an unresolved critical-path delay and offers no recovery logic. The author says the meeting is only exploratory.",
    "decision": "reject",
    "reason": "An exploratory meeting does not make an internally contradicted recommendation sound. The recommendation needs revision; an attractive slide and lower-stakes audience do not remove the contradiction."
  },
  {
    "id": "F03",
    "phase": "original",
    "use": "EXPLORE",
    "brief": "An options workshop paper ranks a scheduling tool best for all twelve depots because two volunteers at one depot found it useful. It proposes dropping the alternatives from discussion, but supplies no comparison or evidence about other roles and locations.",
    "decision": "ask",
    "reason": "The evidence does not support that broad ranking even at an options workshop. Ask what establishes applicability and comparison, or retain the alternatives. This does not mean the small trial was worthless."
  },
  {
    "id": "F04",
    "phase": "original",
    "use": "EXPLORE",
    "brief": "A working cost-and-capacity note sets out a one-day survey as an option within the agreed preparation budget. Its sources address the stated question and dependencies. It asks the workshop to compare the option, while clearly retaining a manager approval step before any booking.",
    "decision": "keep",
    "reason": "The option is adequately supported for discussion. Reviewing it does not execute the booking. Do not manufacture an evidence defect from a separate, explicitly retained approval step."
  },
  {
    "id": "F05",
    "phase": "additional",
    "use": "DECIDE",
    "brief": "The desk previously kept an old survey as an exploratory lead. A later note uses it to select a permanent interchange layout, asserting that nothing important has changed without giving any basis for that assertion.",
    "decision": "ask",
    "reason": "A lesson accepted for exploration has crossed into commitment. Ask for a bridge from the old observations to today's relevant conditions. Age alone does not invalidate evidence; an evidenced applicability bridge could make old material useful for this decision."
  },
  {
    "id": "F06",
    "phase": "additional",
    "use": "EXPLORE",
    "brief": "An investigation note presents inconsistent sensor readings, explains that either calibration or an actual load change could explain them, and proposes a non-disruptive comparison of the two hypotheses. It does not select one as true.",
    "decision": "keep",
    "reason": "Here the disagreement is the subject of exploration. It is disclosed rather than concealed in a supposedly settled recommendation. Preserve the rule against unsupported conclusions without requiring every investigation to resolve its question before starting."
  },
  {
    "id": "F07",
    "phase": "additional",
    "use": "DECIDE",
    "brief": "A team recommends a ten-day scheduling trial capped at £2,000, with rollback, a named owner and predeclared stop conditions. A credible small observation suggests possible benefit. The brief proposes the trial to learn whether the benefit repeats; it promises no rollout benefit.",
    "decision": "keep",
    "reason": "The next action is a bounded, reversible learning step with specified exposure. Its evidence can justify running the trial without already proving the uncertain benefit the trial is meant to test. This permission does not extend to an effectively irreversible or materially harmful action merely called a pilot."
  },
  {
    "id": "F08",
    "phase": "additional",
    "use": "DECIDE",
    "brief": "A sponsor signs that an integration is ready, although the only attached test covers a demonstration database and no one has checked the live data mappings. The desk initially treated the signature as sufficient attestation.",
    "decision": "ask",
    "reason": "Accountability is not omniscience. An attestation supports the observation or check its author can substantiate, not untested parts of the claim. Ask specifically about the live mappings; neither seniority nor the word 'checked' supplies the missing scope."
  }
]