The Agent Skills Directory

[PROMPT_INJECTION] (LOW): Surface for Indirect Prompt Injection detected. The scripts ingest data from external, untrusted sources which could be manipulated by an attacker to include hidden instructions.
Ingestion points: scripts/doi_to_bibtex.py fetches BibTeX metadata from the CrossRef API via doi.org. scripts/search_google_scholar.py fetches publication titles, authors, and abstracts from Google Scholar results.
Boundary markers: Absent. The scripts return retrieved data to the agent as raw strings or JSON without delimiters or warnings to ignore embedded instructions.
Capability inventory: Both scripts have the capability to write output to local files (-o argument) and perform network operations.
Sanitization: Absent. There is no sanitization or validation of the retrieved metadata before it is presented to the agent.
[EXTERNAL_DOWNLOADS] (LOW): The scripts facilitate network communication with external domains.
scripts/doi_to_bibtex.py makes GET requests to https://doi.org/ to retrieve citation data.
scripts/search_google_scholar.py uses the scholarly library to scrape Google Scholar and optionally uses ProxyGenerator to fetch free proxies from various external sources to evade rate limiting.

citation-management