literature-review-agent

To automate literature search and BibTeX generation, this skill performs parallel web searches for candidate papers based on your outline directives. It then verifies the papers via Semantic Scholar and executes Python scripts to deduplicate entries and build a formatted bibliography for your research draft.

The best way to verify candidate papers during citation mining is using the Semantic Scholar API with fuzzy title matching and temporal cutoffs. This approach ensures discovered references are accurate, deduplicated, and formatted correctly before populating your introduction and related work sections.

Yes, you need a web search tool to use this literature review agent. The skill requires web access to find candidate papers, the Semantic Scholar API for verification, and a Python environment to execute provided scripts for deduplication, caching, and BibTeX generation.

Yes, you can generate a bibliography and draft related work sections from an outline. The skill uses your outline and conference guidelines to direct parallel web searches, verify sources via Semantic Scholar, and populate introduction sections with a citation pool enforcing over 90% coverage.

Yes, the Semantic Scholar verification process handles duplicate academic references. It uses fuzzy title matching with a temporal cutoff and executes Python scripts specifically designed for deduplication, ensuring your final bibliography contains only unique, accurate citations.