Outlining : IMRAD structure, claims, and the interview

This is the spec phase. Its output is the input to drafting. Nothing is written as manuscript prose here. Three deliverables:

1. .writing/outline.md, IMRAD structure with 3-7 key claims per section.
2. .writing/claims/section_<NN>_<slug>.md, one file per section, a YAML list of claim stubs prose will later bind to.
3. .writing/metadata.yaml, authors, availability, writing_profile, and reporting fields filled.

The outline is a claim spine, not a table of contents. Every bullet is an assertion the paper will make ("We show X"), not a topic ("Overview of methods"). The claim-first discipline only works if outlining produces real claims.

Two load-bearing contracts

• Filename stems. Sections whose stem ends in _abstract, _references, or _acknowledgments are exempt from paragraph claim-tags. Every other section's .tex file must carry % claim: id-tagged prose. If you add a new exempt section, treat its slug like abstract/references. The contract covers .tex files, not .md.
• Abstract is citation-free. Any _abstract stem emits no \cite/\citep/\citet/\nocite/any \*cite* command and no % claim: tag, it summarizes the paper's own claims in prose. It still carries BPMRC structural tags. Do not create claims/section_00_abstract.md.

Optional term-definition-before-use: copy templates/glossary.md to .writing/glossary.md to activate % define:/% use: ordering checks; add a glossary entry when you add the term's claim stub.

Process