Hypothesis Versioning

Hypothesis Versioning (Step 10)

Save hypotheses with full provenance and manage their lifecycle through iterative refinement.

Instructions

When this skill is invoked:

1. Save artifacts from the current pipeline run:

   cd D:/Claude && python -c "
   from target_to_hypothesis.skills.versioning import save_hypotheses, save_target, save_papers, save_summary
   # save_target(target, run_id)
   # save_papers(paper_candidates, run_id)
   # save_summary(synthesis, run_id, 'synthesis')
   # save_hypotheses(ranked, target, run_id, evidence_summary={...})
   "
   

2. Artifact structure saved to D:/Claude/target_to_hypothesis/artifacts/:

   artifacts/
   ├── targets/       # Normalized target JSON
   ├── papers/        # Retrieved paper metadata
   ├── summaries/     # Synthesis and evidence reports
   └── hypotheses/    # Versioned hypothesis records
       └── {run_id}/
           ├── v001.json    # Version 1 (full hypothesis set)
           ├── v001.md      # Human-readable markdown
           └── log.jsonl    # Append-only status log