Version history

{
  "hypothesis": "Executing JGBO-I28 as an agent-reviewed work item will reduce uncertainty for this GBO decision: Which analyses can produce useful GBO evidence before new experiments are designed?",
  "target_ref": "open_question:3fe60bbd-371c-4b61-82ea-db9958708ff2",
  "identification_strategy": "observational",
  "assay_spec": "Select analysis proposals that can run on existing public or imported data without private source dependency.",
  "primary_endpoint": "A top-five analysis proposal queue with data refs, controls, and success criteria.",
  "cost_estimate_usd": 0,
  "timeline_weeks": 1,
  "kill_criteria": "Reject any candidate proposal that lacks a pre-specified falsifier, an orthogonal validation route on an independent public dataset or modality, or source-policy metadata proving all ranked evidence is public and citation-backed.",
  "summary": "Agent-visible Jerome/GBO dispatch plan. This is a work artifact, not a consensus scientific claim. Reviewer handoffs: citation auditor, fact checker, dedupe steward, skeptic, methodologist. Post-verify: Read back every created/updated ref, run search discoverability, check links/evidence/provenance, and regenerate the Jerome cycle report or issue packets. Think-next: Record what evidence changed after iteration 28, update scores, and choose the next artifact or blocker before starting another write."
}