Xiang Li
Community@coollx
Xiang Li (coollx) maintains a research-assistant skill suite covering academic paper writing, scientific figure visualization, literature management, and experiment lifecycle orchestration.
Agent Skills by Xiang Li
Showing 12 vetted skills indexed across 1 GitHub repositories.
coollxscipilot-figure-skill
Profiles datasets and renders publication-grade scientific figures with matplotlib, seaborn, and plotly.
coollxpress-conference-principle
Guides editorial selection and framing decisions when writing and revising research papers.
coollxdrawio-diagram-builder
Create and iteratively refine editable draw.io XML diagrams using browser screenshot feedback loops.
coollxacademic-humanizer
Rewrites AI-assisted academic drafts and grant proposals to remove AI tells while preserving claims, numbers, and citations.
coollxstyle-synthesis
Synthesizes paper-specific style-guide rules from a corpus of reference academic works.
coollxresearch-paper-writing
Rewrites ML/CV/NLP research paper sections with structured templates and adversarial self-review.
coollxref
Adds, analyzes, and synthesizes reference papers and their code repositories into a structured refs/ library.
coollxanalysis
Produces executed Jupyter notebook reports analyzing experiment runs, datasets, and system behavior.
coollxmeta
Revises the research harness workflow rules, hooks, and skills through an approval-gated process.
coollxtask
Plans and executes research tasks through structured T-NNN task files and spine commits.
coollxdeepxiv-cli
Search and read arXiv and PMC academic papers via the deepxiv command-line interface.
coollxexperiment
Plans and executes research experiment runs with numbered run folders and launch commands.
Frequently Asked Questions About Xiang Li
FAQPage SchemaWhat tasks can I accomplish with coollx's skill suite?▼
You can draft and revise research papers, humanize AI-assisted academic prose, enforce editorial selection principles, generate publication-grade data figures with matplotlib/seaborn/plotly, build editable draw.io diagrams, search arXiv/PMC papers, manage reference libraries, and plan or execute experiments and tasks.
Who is the target user for these skills?▼
ML/CV/NLP researchers, graduate students, and academics preparing journal or conference submissions. The suite also serves grant writers editing NSF Project Descriptions or NIH Specific Aims, and Chinese-language authors needing bilingual figure typesetting for domestic core journals.
How do the figure and diagram skills work in practice?▼
scipilot-figure-skill profiles your data, recommends chart types, blocks common visualization errors, and renders with a render-preview-self-check loop. drawio-diagram-builder creates and iteratively refines .drawio XML diagrams from prompts, papers, screenshots, or existing files with layout fixes.
Are these skills open source and what do they cost?▼
The academic-humanizer skill is explicitly MIT-licensed and free to use. Other skills in the manifest do not declare licenses, so users should check the individual repository folders before redistribution or commercial deployment.
What prerequisites and dependencies are required?▼
Skills target agent runtimes including claude-code, codex, morphmind, opencode, and langchain. Figure generation requires matplotlib, seaborn, SciencePlots, and plotly; Chinese rendering needs Noto Sans CJK, Source Han Sans, or SimHei fonts. Reference management accepts PDFs, arXiv IDs, or URLs.