literaturelisted

# Literature Discovery Search academic databases, curate relevant papers, and populate a project's `docs/literature.md`, bib file, and PDF collection. ## Arguments - `/literature [project-slug]` — run full pipeline for a project - `/literature [project-slug] --refresh` — re-run searches, merge with existing candidates - `/literature [project-slug] --fetch-pdfs` — only download PDFs for already-curated papers - `/literature [project-slug] --cite-graph DOI` — expand candidates via citation graph of a specific paper - `/literature [project-slug] --update <citekey> [--artifact <build/path>]` — **surgical single-entry edit**. Update one literature entry's annotation (key findings, relevance note, supersession marker) when a `/next` iteration produced a result that connects to or contradicts that paper. Reads only `CLAUDE.md`, `literature.md`, `paper/*.bib`, and the optional triggering artifact. Does not re-run searches, does not re-walk the bibliography, does not re-curate. Use from `/next` step 5 when a single citation needs a note added or revised. The `<citekey>` is the BibTeX key as used in `literature.md` and `paper/*.bib`. If no project slug is given, infer from the current working directory. ## Pipeline overview ``` 1. Read context → 2. Generate queries → 3. API search ↓ 4. First curation (Claude) → 5. Citation graph on key papers → 6. Second curation