What problem does it solve?
This Skill solves the challenge of selecting a defensible, fair, and reproducible set of experimental baselines for research papers, ensuring your comparisons stand up to rigorous peer review.
Core Features & Use Cases
- GitHub Reproducibility Gate: Automatically filters out papers that lack usable, non-empty, or runnable code repositories.
- Venue-Aware Recommendations: Tailors baseline sets based on the specific target conference (e.g., ICML, NeurIPS, CVPR) and your available compute budget.
- Reviewer-Risk Audit: Identifies missing classic anchors, recent SOTA, or simple baselines that reviewers are likely to demand.
- Use Case: If you are preparing a submission for CVPR 2027 on 3D detection, use this Skill to generate a defensive baseline set that includes official benchmark methods, recent SOTA, and necessary ablations, while excluding non-reproducible papers.
Quick Start
Use the baseline-selector skill to choose baselines for my research idea on long-context multimodal retrieval targeting AAAI 2027 with a compute budget of 4x A100.