academic-literature-mining - Build and query reproducible scholarly corpora
Build citation-complete scholarly corpora by discovering, screening, acquiring, preserving, indexing, and querying literature with isolated project workspaces.
Tags
Updated: 2026-10-16Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Discover scholarly literature
- Screen candidate records
- Resolve scholarly identifiers
- Verify publication versions
- Acquire authorized full texts
- Preserve PDFs and provenance
- Extract native page text
- Render complete pages
- Embed and rerank pages
- Index vectors in Qdrant
- Query corpus evidence
- Export citation formats
- Audit sources before writing
Inputs
- Research question
- Search concepts
- Inclusion criteria
- Exclusion criteria
- Publication window
- Accepted work types
- Publication identifiers
- Authorized full-text URLs
- User-supplied PDFs
- Project slug
- API credentials
- Research plan
Outputs
- Isolated project workspace
- Preserved source PDFs
- Citation records
- Provenance records
- SQLite corpus state
- Qdrant vector points
- Page-level evidence
- CSL JSON exports
- BibTeX exports
- RIS exports
- Source audit results
Requirements
- Complete canonical Git checkout
- Rust litmine CLI
- Locked Cargo dependencies
- Docker Compose or native Rust
- PDFium
- Qdrant
- SQLite
- NVIDIA Embed VL model
- NVIDIA Rerank VL model
- Authorized access credentials
- User authorization for paywalled access