The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/literature_review.py uses subprocess.run() to invoke sibling Python scripts like scholar_search.py and arxiv_search.py. This orchestration is performed safely by passing arguments as a list, which prevents shell injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes paper abstracts and full-text content from external academic sources such as arXiv and Semantic Scholar. While these are ingestion points for untrusted data, the risk is minimal given the peer-reviewed nature of the sources.
[EXTERNAL_DOWNLOADS]: The skill interacts with public academic APIs including OpenAlex, Semantic Scholar, and arXiv. These are well-known academic resources, and the skill explicitly supports polite API usage via the OPENALEX_EMAIL environment variable.

literature-review