semantic-scholar-deep

To build a citation graph for a literature review, the skill orchestrates Semantic Scholar discovery and multi-hop graph traversal. It batches metadata retrieval and applies Python scripts to map forward citations and backward references around your seed paper, producing a concise ranked report.

Yes, you can run a state-of-the-art survey and novelty check using Semantic Scholar. The skill delegates multi-step literature reviews to a bundled subagent, isolating tokens to traverse citation graphs and snippet search the corpus without cluttering your main context.

The best way to fetch batch lookups and multi-hop references is using the bundled Python scripts. The skill utilizes ss_client.py for fast metadata retrieval and citation_graph.py to orchestrate batching and graph traversal, returning compact reports instead of raw API data.

This Semantic Scholar deep research workflow requires no external dependencies but needs access to the bundled Python scripts. It orchestrates Exa MCP discovery and the semantic-scholar-lookup integration internally to fill in-depth reference gaps.

Semantic Scholar often lacks in-depth forward citations and backward references due to standard API limits. This skill fills those gaps by orchestrating Exa MCP discovery with multi-hop citation graph traversal and batch lookups to generate comprehensive reference data.