Paper-to-HTML Skill

paper-to-html is a Codex skill for turning academic papers, arXiv pages, local PDFs, project pages, or pasted paper text into a clear Chinese paper note with:

• a responsive HTML file
• a matching browser-rendered PDF
• a structured explanation of motivation, method pipeline, network/module roles, experiments, limitations, and code status

The skill does not depend on any other custom local skill. Everything required by the skill is contained in this directory.

Directory Layout

paper-to-html/
├── SKILL.md
├── README.md
├── agents/
│   └── openai.yaml
├── references/
│   └── minimal-note-style.md
└── scripts/
    └── export_pdf.py

Installation

Clone this repository into your Codex skills directory:

Please install Thomasnxz/Paper-to-HTML for me.
GitHub: https://github.com/Thomasnxz/Paper-to-HTML.git

If you install it somewhere else, make sure the final directory name is paper-to-html and Codex can discover that skill directory.

Usage

In Codex, call the skill with a paper URL, PDF path, project page, or pasted content:

$paper-to-html https://arxiv.org/abs/1706.03762

By default, outputs are written under paper_outputs/ in the current workspace:

paper_outputs/<paper-name>.html
paper_outputs/<paper-name>.pdf
paper_outputs/<paper-name>_assets/

PDF Export

The skill uses the bundled script:

python "${CODEX_HOME:-$HOME/.codex}/skills/paper-to-html/scripts/export_pdf.py" /absolute/path/note.html

The script writes a same-name PDF by default:

/absolute/path/note.html
/absolute/path/note.pdf

You can also pass an explicit PDF path:

python "${CODEX_HOME:-$HOME/.codex}/skills/paper-to-html/scripts/export_pdf.py" /absolute/path/note.html --pdf /absolute/path/out.pdf