expand-references

What problem does it solve?

Finding the next papers to read after you already have a few key “anchor” works can be slow and unfocused.

Core Features & Use Cases

• Curated neighborhood expansion: Expands one to three seed papers into structured follow-ups, including the closest neighbors, bridging papers, foundational work, methodological papers, recent work, and surveys/benchmarks.
• Field- and venue-aware scoring: Ranks candidates using similarity to the seeds, impact signals, recency, venue novelty, and textual overlap.
• Cluster steering with negatives: Lets you push the expansion away from an unwanted subtopic by specifying negative papers.
• Use Case: You have three anchor papers in retrieval-augmented generation and want a prioritized sequence of foundational background, methods, and the most useful recent surveys without manually searching.

Quick Start

Run the skill by providing one to three seed paper titles (quoted if multi-word) and optionally add negative seeds, then read the structured buckets from the returned JSON fields.