Verified Agent Skill record

Huggingface Paper Publisher

This skill provides tools to publish, manage, and link research papers on the Hugging Face Hub. It supports indexing papers from arXiv, linking papers to models/datasets/spaces via YAML metadata, claiming authorship, and generating professional markdown-based research articles.

Tasks

Index a paper from arXiv on Hugging Face using `uv run scripts/paper_manager.py index --arxiv-id "<arxiv-id>`"
Link a paper to a model or dataset repository by running `uv run scripts/paper_manager.py link --repo-id "<repo-id>" --repo-type model --arxiv-id "<arxiv-id>`"
Claim authorship of a paper using `uv run scripts/paper_manager.py claim --arxiv-id "<arxiv-id>" --email "<institution-email>`"
Generate a research article from a template using `uv run scripts/paper_manager.py create --template modern --title "<title>" --output paper.md`

Inputs

arXiv IDs (e.g., 2301.12345)
Hugging Face repository IDs (e.g., username/model-name)
Repository types (model, dataset, space)
Author names, titles, and abstracts
BibTeX citation text

Outputs

Indexed Paper Pages on Hugging Face Hub
Markdown research articles (paper.md)
HTML research articles (paper.html)
YAML metadata updates in model and dataset README files
BibTeX citations and JSON paper metadata

Limitations and checks

arXiv ID must exist or be indexed for paper operations
Authorship claims require admin review and verification via institutional email
Rate limiting may occur if too many API requests are made in a short time
Visibility toggling only applies to papers where authorship is already claimed and verified
Verify that the `HF_TOKEN` environment variable is set with an access token that has write permissions before running scripts
Run `uv run scripts/paper_manager.py check --arxiv-id "<id>"` to confirm whether a paper exists on the Hub before indexing or linking
Ensure the README YAML frontmatter in the target repository is properly formatted to allow the Hub to extract arXiv IDs and auto-generate tags