The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and process research paper content from Hugging Face and ArXiv in markdown format. This external data represents an untrusted source that could contain hidden instructions intended to override agent behavior during summarization or analysis tasks.
Ingestion points: The agent fetches paper content via curl "https://huggingface.co/papers/{PAPER_ID}.md" as described in SKILL.md.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' prompts when handling the retrieved content.
Capability inventory: The agent utilizes network tools (curl) and standard text processing capabilities.
Sanitization: No explicit sanitization or filtering of the fetched markdown is prescribed.
[COMMAND_EXECUTION]: The skill provides documentation for using curl commands to query Hugging Face APIs for metadata, search results, and to perform authenticated operations like claiming authorship or indexing papers.
[EXTERNAL_DOWNLOADS]: The skill facilitates data retrieval from huggingface.co and arxiv.org. These are well-known services for the AI research community, and the skill's interactions with them for metadata and document fetching are consistent with its stated purpose.

huggingface-papers