What problem does it solve?
Manually following a paper's references and citing papers is slow and error-prone, and naive traversal explodes exponentially. This Skill automates citation graph expansion through OpenAlex while enforcing relevance filtering, deduplication, and depth limits so literature reviews stay focused.
Core Features & Use Cases
- Bidirectional Citation Traversal: Follow a seed paper's references (backward) and citing papers (forward) in one call via the rp_cite tool, with abstracts returned inline for immediate scoring.
- Relevance-Gated Expansion: Only expand seeds scoring >= 7 and enqueue citations scoring >= 5, with depth capped at 2 levels and checkpoints every 50 papers.
- Deduplication & Graph Tracking: Skip already-reviewed papers via papers-reviewed.json and record all relationships in citations/citation-graph.json.
- Use Case: After finding a highly relevant paper on multi-agent debate, traverse its references and citing papers to build a related-work corpus, scoring each candidate before adding it to the review queue.
Quick Start
Use the traversing-citations skill to expand from this seed DOI in both directions, filter candidates by relevance, and record the results in the citation graph.
