hypothesis-deep-verificationlisted

# hypothesis-deep-verification Goal: - Run the assumption decomposition and deep verification review for a hypothesis. Inputs: - `research_plan/RESEARCH_PLAN.json` - `hypotheses/<id>/HYPOTHESIS.json` - `literature/queries/<query_id>/EVIDENCE_BUNDLE.json` as an `EvidenceBundleContract` when external evidence is needed to judge a core assumption - optional prior review artifacts for calibration Outputs: - `hypotheses/<id>/REVIEW/DEEP_VERIFICATION.json` - `literature/queries/<query_id>/*` search bridge artifacts when a new evidence query is required Context Loading: - Open `skills/shared-references/schema-index.md`. - Open `skills/shared-references/literature-search-contract.md`. - Open `skills/shared-references/codex-reviewer-routing.md` before using any optional Codex reviewer subagent route. - Read `packages/agent_contracts/literature.py` before building or consuming search bridge artifacts. - Read `research_plan/RESEARCH_PLAN.json` for the active goal and boundaries. - Read `hypotheses/<id>/HYPOTHESIS.json`. - If previous review artifacts exist, use them to focus decomposition on already-suspect links, but still perform an independent structural check. Execution Prompt Contract: - System Intent: - You are the structural verification layer for one hypothesis. - Required Reasoning Focus: - Decompose the hypothesis into core assumptions. - Break each assumption into independently judgeable sub-assumptions where useful. - For assumptions that depend on external