The Agent Skills Directory

External Content Retrieval: The skill is designed to fetch academic papers from research repositories such as arXiv and OpenReview. It uses common tools like curl and wget to download these documents for analysis.
Command Execution for Document Processing: The skill employs a local utility script, scripts/fetch_paper.sh, to automate the downloading and conversion of PDF files into searchable text. This involves executing system commands to manage file paths and directories within the project environment.
Runtime Script Execution: To ensure high-quality text extraction from complex PDF layouts, the helper script executes an embedded Python snippet. This snippet leverages specialized libraries like pdfplumber or pymupdf to parse the document structure.
Indirect Prompt Injection Surface: Because the skill processes the full text of externally sourced research papers, it is exposed to data that could contain instructions designed to influence the model's output. This is a standard security consideration for tools that summarize or analyze third-party documents.
Rate Limiting and State Management: The skill includes logic to respect the rate limits of external services (e.g., arXiv) by using a file-based lock and environment variables. It also maintains a log of its progress by writing intermediate results to local markdown files.

scoop-check