The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the retrieval and processing of scholarly works from multiple external APIs (arXiv, PubMed, bioRxiv, etc.), representing a surface for indirect prompt injection via academic metadata.
Ingestion points: External data enters the agent context through API responses containing paper titles, abstracts, and full-text content in files like arxiv-api/SKILL.md, pubmed-api/SKILL.md, and arxiv-paper-processor/SKILL.md.
Boundary markers: The provided implementation examples do not include explicit prompt delimiters or instructions to ignore embedded commands within the retrieved research data.
Capability inventory: The skill includes capabilities for network operations (requests, curl), file system writes (arxiv-batch-reporting/SKILL.md), and the execution of external scripts (chatpaper-guide/SKILL.md).
Sanitization: There is no evidence of content sanitization or validation for the text retrieved from academic databases before it is displayed or processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill documents the installation of third-party research tools and libraries.
Evidence: Instructions for pip install arxiv-cli-tools in arxiv-cli-tools/SKILL.md and git clone https://github.com/kaixindelele/ChatPaper.git in chatpaper-guide/SKILL.md.
Context: These are standard installation procedures for well-known academic utilities from reputable sources.
[COMMAND_EXECUTION]: The skill provides instructions for executing various command-line tools to process and analyze retrieved papers.
Evidence: Usage of docker run for the GROBID service in arxiv-paper-processor/SKILL.md and shell-based execution of the arxiv-cli and chat_paper.py tools in their respective guides.

search-skills