Academic Search

What it does

This skill provides robust academic search and metadata-extraction capabilities. It helps an agent locate papers, gather structured metadata (title, authors, year, venue, citation counts, DOI/arXiv IDs), fetch PDFs or BibTeX entries, and combine results from multiple platforms into a deduplicated schema. The skill includes guidance on selecting the right source (arXiv, Semantic Scholar, PubMed, Papers With Code, ACM DL, IEEE Xplore, CNKI) and when to use browser automation (CDP) for sites without APIs like Google Scholar or CNKI.

When to use it

Use this skill when a user asks to find papers, build a literature review, retrieve citation counts or citation graphs, export BibTeX, fetch PDFs, or assemble an annotated bibliography. It triggers on queries about papers, venues, citations, authors, arXiv/DOI lookups, and when links to known academic domains appear.

What's included

• Scripts: repository provides scripts for checks and CDP helpers (has_scripts=true).
• References: platform patterns and metadata schema guidance are included in references (has_references=true).
• Instructions: procedural guidance on a two-pass search strategy (light scan then deep-dive), platform selection matrix, PDF/BibTeX retrieval order, CDP usage templates, and failure signals for robust retries.

Compatible agents

Best suited for agents with HTTP and headless-browser capabilities (Claude Code / agents that can run Node/Curl and CDP). The skill assumes access to network APIs and optionally a CDP proxy for Google Scholar and CNKI.