Rubric Rumble: extract model facts from supplied documentation

Rubric Rumble: extract model facts from supplied documentation A sequence diagram generated by Archify. Desktop explicit request: reserve output first Supply text; URL is provenance, not fetched Validate input, model settings and size limits One extraction request to researcher Complete nonempty response and usage Strict JSON; target identity; field types Known facts require source quotes; unknowns null Catalog + source hash + researcher receipt Save model-research.json; show supported limits Research tab / API · Sequence participant Research tab / API Research evidence · Sequence participant Research evidence research_model · Sequence participant research_model Researcher model · Sequence participant Researcher model Quote validation · Sequence participant Quote validation Legend return default message

Researcher and target are different roles

  • • Only the configured researcher model is called. The target ID is what the supplied document is about; this operation does not probe the target.

Textual provenance still needs judgment

  • • Quotes establish textual provenance, not correctness. Review aliases; task limits are not auto-edited. Recording time is not document publication time.