Dev & Engineering hugging-face-hubarxivpaper-publishingmodel-cardsdataset-cardscitation-managementmarkdown-templatesyaml-metadata

HF Paper Publisher

Index arXiv papers on Hugging Face Hub, link them to model and dataset cards, and manage authorship and citations in one place.

FollowSkills review · FSRS-2.0
Not recommended
45/ 100 5-point scale 2.3 / 5
1 2 3 4 5 6
1Trust12 / 25 · 2.4/5

The script only requests HF_TOKEN and performs declared operations (README writes, arXiv/HF queries); no covert exfiltration or sensitive-data misuse. However, the link command commits directly to remote repos with no confirmation, dry-run, backup or rollback; --create-pr is silently unimplemented and falls back to direct commits; docs advise writing the token to a .env file without leakage warnings. Deducted for missing write confirmation, isolation and recovery.

2Reliability7 / 20 · 1.8/5

Core paths (check/index/create) are self-consistent with readable errors, but several commands advertised in SKILL.md (convert, list-my-papers, toggle-visibility, check-authorship, validate) do not exist in paper_manager.py; search just prints 'coming soon'; --create-pr silently does nothing; citation year is guessed from the arXiv ID prefix (wrong for older papers); the injected '## Paper' section is inserted mechanically and may break layouts. Deducted for doc-implementation mismatch, no tests, thin edge handling.

3Adaptability7 / 15 · 2.3/5

The target scenario (publishing/linking arXiv papers on HF for engineers/researchers) is clear with precise semantic triggers and thorough YAML/link format docs; but capability boundaries are undeclared (which commands actually work), and all core function depends on huggingface.co and arxiv.org, raising mainland-China reachability concerns; no Chinese support. Deducted for overseas-service dependence without disclosure and missing boundary statements.

4Convention9 / 15 · 3.0/5

SKILL.md is well layered (overview, dependencies, usage, troubleshooting, references), versioned 1.0.0, MIT licensed, with templates and examples separated; but there is no changelog, no stated maintenance ownership, and the documented capability exceeds the implementation without clear labeling. Deducted for incomplete version governance and doc-implementation alignment.

5Effectiveness6 / 15 · 2.0/5

Working paths (template creation, check, bibtex citation) offer real marginal value and the four templates are directly usable; but link rewrites remote READMEs mechanically, the 'expected outputs' in examples are fabricated demos rather than verified results, and missing advertised commands force user-side debugging. Deducted for unverified output usability and a shrunken command surface requiring rework.

6Verifiability4 / 10 · 2.0/5

All source, templates and examples are auditable and facts are mostly separated from inference; but the 'Expected output' blocks are authored demos, there is no test suite and no CI covering this skill's key paths, and network behavior cannot be reproduced in a static read. Deducted for lack of third-party execution evidence and independently unverifiable sample outputs.

Evidence confidence:Low Reviewed Sep 09, 2026 Reviewed revision 2c9b106168d4
Before you use it
  • The link command commits README changes directly to remote HF repos with no confirmation or rollback; back up manually or verify read-only first on important repositories.
  • Commands advertised in SKILL.md (convert, list-my-papers, toggle-visibility, check-authorship, validate) do not exist in the script; search is a stub and --create-pr silently does nothing. Cross-check the script's actual argparse subcommands before use.
  • Citation years are inferred from the first two digits of the arXiv ID and will be wrong for pre-2000 papers; manually verify generated citations.
  • Core functionality depends on huggingface.co and arxiv.org, which may be unreachable from mainland-China networks; the documented .env approach for HF_TOKEN carries leakage risk—prefer system-level secret management.
  • This is a static source review with no execution; runtime conclusions carry limited confidence.
See the full review method →

What does this skill do, and when should you use it?

This skill gives AI researchers and engineers a set of Python scripts covering the Hugging Face Hub paper ecosystem. It indexes paper pages from arXiv IDs, writes paper links into model/dataset card YAML metadata (the Hub auto-generates arxiv: tags), initiates authorship claims, and generates Markdown research articles from four templates. It requires an HF_TOKEN with write access. It is one skill inside the patchy631/ai-engineering-hub repository, which is MIT-licensed.

Runs via scripts/paper_manager.py with subcommands: index (index a paper page by arXiv ID), check (see if already indexed), link (insert paper links and citations into model/dataset/Space READMEs, supporting multiple papers and custom BibTeX), claim and check-authorship (authorship verification), toggle-visibility and list-my-papers (control profile display), create (generate articles from standard/modern/arxiv/ml-report templates), convert (Markdown to HTML), plus search, info, citation, and validate. It calls the Hugging Face Hub API, reads and writes README YAML frontmatter, and supports options like --create-pr.

  1. A researcher who just submitted a paper to arXiv and wants a Hugging Face paper page immediately
  2. A model author who wants to properly cite the method paper in a model card so the Hub auto-tags it
  3. A paper author claiming authorship on a HF paper page and controlling its profile visibility
  4. A team linking one paper to multiple models, datasets, and Spaces at once
  5. Anyone needing a well-structured Markdown research report converted to shareable HTML

What are this skill's strengths and limitations?

Pros
  • Covers the full publishing workflow: indexing, linking, authorship, visibility, and article generation
  • CLI scripts are reusable directly and exposed via a PaperManager Python class
  • Very thorough documentation: full command reference, YAML examples, error handling, and workflow examples
  • Supports batch linking of multiple papers and custom citation text
Limitations
  • Tied to the Hugging Face ecosystem; non-arXiv venues are unsupported and only listed as future plans
  • Authorship claims depend on manual HF admin review with unpredictable timing
  • No test suite or standalone release history is documented (skill version is labeled 1.0.0)
  • Requires holding a write-access HF_TOKEN, adding credential-management overhead

How do you install this skill?

The skill lives at hugging-face-skills/skills/hugging-face-paper-publisher/ in patchy631/ai-engineering-hub. Install dependencies: uv add huggingface_hub pyyaml requests markdown python-dotenv, set the HF_TOKEN environment variable (write access required), and activate the venv: source .venv/bin/activate. Drop the skill folder into your Agent Skills-compatible client's skills directory; the exact path is not documented in the source.

How do you use this skill?

Typical commands: python scripts/paper_manager.py index --arxiv-id "2301.12345" to index a paper; python scripts/paper_manager.py link --repo-id "username/model-name" --repo-type "model" --arxiv-id "2301.12345" to link it to a model card; python scripts/paper_manager.py create --template "modern" --title "Paper Title" --output paper.md to generate an article; python scripts/paper_manager.py claim --arxiv-id "..." --email "..." to claim authorship. In an Agent Skills client these can be triggered via natural language.

How does this skill compare with similar options?

The skill docs mention tfrere's research-article-template (a Hugging Face Space) as complementary: that template is for writing research articles, while this skill handles indexing, linking, and metadata on the Hub. No other competitors are named in the source.

FAQ

What permissions and costs are involved?
You need a Hugging Face account and an HF_TOKEN with write access; the Hub is free and the skill is MIT-licensed, so there is no additional cost.
How is a linked paper discovered?
The Hub extracts the arXiv ID from the README link and auto-generates an arxiv:<PAPER_ID> tag; clicking the tag opens the Paper Page, which lists all models and datasets citing the paper.
What if my authorship claim fails?
Common causes are an email that doesn't match the paper's author records or a claim already made by someone else; use an institutional email, wait for admin review, and contact HF support with proof if needed.
Does it support non-arXiv papers?
Currently only arXiv IDs are supported; conference proceedings and journal papers are listed as planned future enhancements.

More skills from this repository

All from patchy631/ai-engineering-hub

Dev & Engineering

Hugging Face Evaluation Manager Skill

Structure benchmark scores into Hugging Face model cards: extract them from READMEs, import from Artificial Analysis, or run your own evals with vLLM/lighteval.

Dev & Engineering

Hugging Face API Tool Builder

Turns Hugging Face API interactions into reusable, pipe-composable command-line scripts instead of one-off fetch code written every time.

Dev & Engineering

Hugging Face CLI Skill

Lets your AI assistant run Hugging Face Hub operations in the terminal — model downloads, uploads, repo management, cache cleanup, and cloud GPU jobs.

Data & Analysis

Hugging Face Datasets Skill

Create, configure, and stream datasets on the Hugging Face Hub, and query, transform, and republish any public dataset with DuckDB SQL.

Dev & Engineering

Hugging Face Jobs Runner Skill

Lets your AI assistant submit any Python workload to Hugging Face's fully managed cloud compute — no local GPU or setup — while safely handling auth, timeouts, and result persistence.

Data & Analysis

Trackio Experiment Tracking Skill

Log metrics during model training, retrieve and analyze them afterward, and sync dashboards to Hugging Face Spaces for real-time monitoring.

Dev & Engineering

HF Model Trainer (TRL on Hugging Face Jobs)

Fine-tune language models with TRL on Hugging Face Jobs cloud GPUs — no local GPU needed — with automatic Hub persistence and GGUF conversion.

Dev & Engineering

Bright Data Web MCP Skill

Reliable web access for MCP-compatible agents: search, scrape, extract structured data, and automate browsers with automatic anti-bot and CAPTCHA bypass.

Dev & Engineering

GRPO Fine-Tune Skill (Qwen3 / Fireworks)

Turn a plain-English task description plus a dataset into a full GRPO reinforcement-learning fine-tuning run on Fireworks-managed GPUs, with no training code to write.

Related skills