Wanghan Xu
Community@black-yt · Shanghai, China
PhD student at Shanghai Jiao Tong University & Visiting Scholar at Stanford University
Agent Skills by Wanghan Xu
Showing 26 vetted skills indexed across 1 GitHub repositories.
frontend-markdown-rendering
Render final assistant responses as sanitized Markdown with KaTeX, Mermaid, and tables.
pdf-parsing
Convert PDF documents into local Markdown and extracted image assets.
ai-conference-paper-writing
Rewrite AI conference paper narratives with aligned thesis and sections.
lab-cluster-1
Manage PJLAB lab cluster SSH, job submission, and service workflows.
codex-history-sync
Synchronize Codex history between .codex directories with dry-run validation.
agents-md-maintenance
Organize AGENTS.md and CLAUDE.md files into entry and local detail documents.
docx-splitting
Split DOCX documents into page-based parts using Word COM automation.
huggingface-dataset-publishing
Publish Hugging Face datasets with image and metadata handling.
latex-compiling
Compile LaTeX sources into PDFs with latexmk and isolated build directories.
context-overlay
Configure and debug an OpenAI-compatible context overlay proxy with YAML rules.
markdown-to-docx
Convert Markdown documents into Word .docx files with embedded images and captions.
llm-deploy-training
Deploy and train LLM and VLM systems with version-aware configuration and validation.
markdown-to-pdf
Convert Markdown documents to PDF with tables, code blocks, and images.
browser-render-visualization
Render web pages in a real browser to diagnose blank screens and runtime errors.
pdf-to-images
Convert PDF pages into PNG or JPEG images with Ghostscript or Poppler.
wsl-bash-windows
Diagnose and guide shell, path, and network execution in Windows and WSL mixed environments.
latex-template-migration
Migrate LaTeX papers between templates while preserving content and submission readiness.
researchharness
Configure and operate ResearchHarness as a tool-using LLM agent runtime.
rebuttal-writing
Draft anonymous AI and ML conference rebuttals from reviewer comments and evidence.
llm-image-generation
Generate images via an OpenAI-compatible LLM gateway and save them locally.
lab-cluster-1-web-portal
Log into the lab-cluster-1 portal and query Prometheus and DCGM metrics.
feishu-bot
Configure, send, and troubleshoot Feishu and Lark bot messages.
overleaf-project-sync
Compare local LaTeX directories with Overleaf project ZIPs and update files.
structai
Build and debug structured LLM workflows with the StructAI Python toolkit.