Resolve the test data a paper names into refs for each workflow input.
Every file beside index.md is one the mold kind declares.
source-specific fields are coherent.
Abstract oracle: the properties any cast of this Mold must satisfy.
eval.md not written yet.
Concrete cases bound to fixtures, run against the eval properties.
scenarios.md not written yet.
1 typed reference; 0 resolver issues.
All on-demand references describe triggers.
Hypothesis references include verification.
Typed references describe what casting consumes, and when the generated skill should load each artifact.
Structured contract copied for validation or lookup.
Test data the paper names, one ref per workflow input (URLs, shapes, hashes). Shared id with [[find-test-data]], which fills only the inputs this Mold leaves unresolved.
Inputs sharing a role are alternatives — one producer satisfies it.
Free-form summary from [[summarize-paper]]; sample data and reference evidence the test fixtures derive from.
Galaxy interface brief from [[freeform-summary-to-galaxy-interface]] pinning the workflow input labels, collection shapes, and datatypes each ref must map onto, in PAPER → GALAXY.
CWL design from [[freeform-summary-to-cwl-design]] naming the workflow input ports each ref must map onto, in PAPER → CWL.
Read the paper free-form summary from summarize-paper and resolve the data the paper names (resolvable URLs, file shapes, expected hashes) onto the workflow inputs the target declares — the Galaxy interface brief’s input labels, or the CWL design’s input ports — as test-data-refs with resolved_by: paper-to-test-data. An input the paper names no data for stays resolved: false with a reason, for find-test-data to search. Expected outputs are the test plan’s to derive, not this Mold’s. Validate the file with foundry validate-test-data-refs before returning.