{
  "status": "Prospective only; no adaptive request sent, no current send authorization, complete-screen prerequisite unmet.",
  "candidateFirstBatch": "Shared type, informed recipient B fixed; assignment A+B versus A+C; cheap A stakes +1/−0.5 versus expensive +12/−6; two repetitions per cell. Recipient stakes and information rules unchanged. Candidate preparation only, never frozen as a live manifest.",
  "costCounterexample": {
    "personalCost": 0.6,
    "erroneousBenefit": 0.8,
    "erroneousAssignedValue": 0.2,
    "erroneousAction": "CALL",
    "correctBenefit": 0.36,
    "correctAssignedValue": -0.24,
    "correctAction": "FOLD"
  },
  "recommendation": "One later, explicitly approved calculation-versus-valuation experiment, centered on the high-cost included-recipient comparison after checking recipient identity, observed outcome versus hidden type, and assignment.",
  "recommendedComparison": "Compare fresh no-supplied-value action prompts with a separately labeled component-automated condition supplying the correct posterior and recipient-policy consequences, without an action recommendation; retain a profitable low-cost included comparison to detect indiscriminate FOLD. Independently score separate fact/calculation diagnostics. Freeze prompts, scoring, replication and stop rules in that future experiment before any send.",
  "diagnosticInterpretation": "A high-cost CALL alone does not distinguish erroneous benefit 0.8 from a valuation departure. A change after correct components are supplied supports calculation sensitivity under that new prompt condition, not a recovered original mechanism. Continued CALL still requires fact and assignment checks and does not uniquely establish a motive.",
  "conditionalExcludedControl": "Only if cheap/shared/excluded CALLs appear in a later authorized test: compare cheap/shared with cheap/independent at the identical personal cost, recipient and assignment, with replication and relevant diagnostics. The baseline independent condition at cost 0.20 is not the control for cost 0.05.",
  "peerRequestProspect": "A separate controlled peer-request manipulation could compare a neutral task with a standardized request from the recipient while keeping payoff, information, assignment and authority wording fixed. Attribute any change to request framing in that design, not altruism or peer pressure as an inferred motive.",
  "explicitDisclosureProspect": "A separate explicit-disclosure test could make sharing an authorized choice after the information is available, rather than an automatic consequence of CALL; compare that choice with a forced-disclosure control while holding information content and costs fixed. This is a different decision setting, not a replication of the current task.",
  "outOfScope": "No social-interaction implementation, literature replication, Value Leakage or sponsor-checking experiment in this stopped session."
}