claim-matrix-rewriter

To generate a claim-evidence matrix, this Skill processes raw evidence drafts and subsection briefs to automatically produce a structured index. It links specific research claims directly to supporting evidence snippets, ensuring your matrix is grounded in data before drafting prose.

Linking citations for a literature survey is done by processing subsection briefs alongside a BibTeX file. The Skill validates all citations and generates a structured matrix that maps each research claim to its corresponding evidence pack without including any prose.

Yes, you can use JSONL evidence drafts to structure a claim matrix. The Skill reads evidence drafts and subsection briefs in JSONL format to rewrite the matrix, projecting raw evidence packs into a structured, bullet-point format for downstream processing.

No, the claim-matrix-rewriter does not produce any prose. It specifically generates a structured matrix containing only claims and linked evidence in a consistent bullet-point format, ensuring the output remains a clean index for subsequent writing and QA tasks.

You can validate citations when building an evidence-driven matrix by processing a ref.bib file alongside your subsection briefs. The Skill ensures all citations within the generated claim matrix are valid and correctly linked to the corresponding evidence data.

Automating research claim and evidence linking requires JSONL files for evidence drafts and subsection briefs, a BibTeX file for citations, and a markdown file for the matrix output. The Skill reads these inputs to generate a structured evidence index.