Skill UI

This skill runs the ScholarEval framework to score academic submissions across problem setting, methodology, data, analysis, writing, and references, providing structured strengths/weaknesses, numeric ratings, and guidance for visualizing the evaluation.

This skill guides users in systematically organizing and documenting a year's worth of past projects into the specific SRED format. It processes scattered evidence—including GitHub PRs, Notion documents, and Linear tickets—to evaluate project suitability, generate comprehensive descriptions, and create structured, standardized reports within Notion. Ideal for formal academic submissions or professional portfolio creation.

This skill automates the scholarly process of refining a generated academic paper. It executes an iterative loop (Review → Fix → Recompile) using advanced LLMs (like GPT-5.5) over multiple rounds. It is designed not for primary research, but specifically for polishing writing quality, correcting theoretical inconsistencies, improving presentation, and enhancing narrative flow after the initial draft is complete. It helps elevate a paper from a draft to a polished submission.

Automates the entire process of converting LaTeX source files into a professional, submission-ready PDF. This tool handles multi-pass compilation (using latexmk), diagnoses common LaTeX errors (such as missing packages, undefined references, and BibTeX syntax errors), and performs post-compilation checks, including page count verification according to specific academic guidelines (ICLR, IEEE, etc.). Ideal for researchers submitting papers.

This skill automates the generation of high-quality, publication-ready figures and comparison tables from raw experimental data (CSV, JSON, etc.). It supports various plot types, including line plots, bar charts, scatter plots, heatmaps, and box plots, ensuring consistent style (DPI 300, PDF format) suitable for academic submissions. Users provide a figure plan and raw data, and the system generates the necessary code to visualize results.

Orchestrates a comprehensive, end-to-end pipeline for academic paper generation. It guides the user from a raw narrative report or research idea to a polished, submission-ready PDF. The process includes structured planning, figure generation (via SVG, Gemini, or Mermaid), LaTeX writing, compilation, and an automated multi-round improvement loop, ensuring the final output adheres to specified venue styles (e.g., ICLR, NeurIPS).

A comprehensive tool for generating publication-quality figures essential for Machine Learning and AI academic papers. It supports two distinct workflows: generating complex system architecture and workflow diagrams using AI (Gemini), and creating precise, data-driven visualizations (line charts, heatmaps, etc.) using standard libraries like matplotlib and seaborn. Use it when preparing any figure for a conference submission.

Access comprehensive LaTeX templates, formatting requirements, and detailed submission guidelines for major academic venues. This skill supports the creation of manuscripts destined for top-tier journals (Nature, Science, Cell), international conferences (NeurIPS, ICML, CVPR), research posters, and major grant proposals (NSF, NIH, DOE). It ensures your work meets the specific formatting standards required by the target publication.

This skill automates the complex process of resubmitting an already polished academic paper to a different conference or journal. It is designed for scenarios where the core content, experiments, and bibliography must remain untouched, but the paper needs micro-edits to accommodate new venue constraints, page limits, or prior reviewer feedback. It ensures the new submission is generated as a separate, sibling directory to maintain the integrity of prior submissions.

A comprehensive orchestrator designed to manage the entire academic research cycle, from initial deep research to final manuscript submission. It guides users through mandatory stages including writing, integrity verification, two-stage peer review, and revision. This skill does not perform the substantive work itself; instead, it tracks state, recommends modes, and dispatches specialized skills (like deep-research and paper formatter) in a reproducible, controlled workflow, ensuring academic rigor at every step.