The Agent Skills Directory

[External Content Sanitization]: The skill processes metadata retrieved from external sources such as arXiv. It includes a dedicated _sanitize_text method in the paper_manager.py script that neutralizes control characters, prevents markdown code fence breakout, and escapes YAML document delimiters. This measure ensures that untrusted content from research papers does not disrupt the repository's configuration or documentation structure.
[Credential Management]: Authentication for repository operations is handled using the HF_TOKEN environment variable. The skill is designed to load this token via the python-dotenv package, which is a standard and recommended practice for managing sensitive API tokens in a development environment without hardcoding them in the source code.
[Integration with Trusted Services]: The tool interacts with well-known and trusted technology platforms, specifically Hugging Face and arXiv. It utilizes the official huggingface_hub Python client and standard API calls to these services, ensuring that data transfers and repository modifications are conducted through established, legitimate channels.

huggingface-paper-publisher