Literature Review

**Judgement:** A strong, domain-appropriate skill for academic literature discovery and review drafting. It is well-suited as a *default* choice whenever a user explicitly requests scholarly papers, DOIs/PMIDs, or help writing a literature review. **Strengths & value:** - Multi-source integration (Semantic Scholar, OpenAlex, Crossref, PubMed) with DOI-based deduplication gives broader coverage and cleaner result sets than single-API tools. - Returns full abstracts for most sources, enabling the model to do more meaningful synthesis rather than shallow keyword matching. - Includes an explicit synthesis workflow (extract → organize → draft), making it easier to produce structured review sections instead of ad-hoc summaries. **Key risks / limitations:** - Dependent on external APIs (rate limits, downtime, changing schemas); failures or partial results may silently degrade coverage. - Crossref entries may lack abstracts; if not handled carefully, the model could over-interpret papers from minimal metadata. - Results are not a substitute for expert judgement, especially for clinical or high-stakes decisions; the assistant must avoid treating retrieved literature as direct medical or legal advice. - Coverage, citation counts, and indexing lag may lead to missing very recent or niche work. **Recommended usage scenarios:** - User asks: “Find papers on X” or “What does the literature say about Y?” and wants references, DOIs/PMIDs, or key summaries. - User needs help drafting or structuring a related-work / background / literature review section with citations. - Tasks requiring identification of influential or recent work (filtering by year or citation counts) rather than just generic web search results. **When to be cautious or consider alternatives:** - Time-critical, clinical, or policy decisions where incomplete coverage or misinterpretation of abstracts could cause harm—use this tool for evidence gathering only, with clear disclaimers. - Situations where only local, proprietary, or non-indexed documents are relevant; this tool only sees what the public APIs expose.