The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/generate_pdf.py uses subprocess.run to execute external system tools pandoc and xelatex. While this is necessary for the skill's primary function of generating formatted PDFs, it involves executing system-level binaries.
[EXTERNAL_DOWNLOADS]: The scripts/verify_citations.py script performs network requests to well-known and trusted academic metadata services, specifically doi.org and api.crossref.org. These connections are used to validate publication DOIs and retrieve bibliographic data.
[PROMPT_INJECTION]: The skill represents an indirect prompt injection surface because it is designed to ingest and process content from external sources such as PubMed, arXiv, and preprint servers. If these sources contain maliciously crafted text (e.g., in abstracts or titles), they could potentially influence the agent's synthesis or analysis.
Ingestion points: External data is ingested through search results processed by scripts/search_databases.py and metadata retrieved by scripts/verify_citations.py.
Boundary markers: While the assets/review_template.md provides document structure, the skill lacks specific delimiters or "ignore previous instructions" safety markers to isolate the agent from instructions embedded within the processed research data.
Capability inventory: The skill has access to Read, Write, Edit, and Bash tools, and it executes system commands via its own Python scripts.
Sanitization: The provided scripts do not show evidence of sanitizing or escaping the text content retrieved from external literature databases before it is presented to the agent for synthesis.

literature-review