citation-management

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting untrusted content from external academic databases and URLs.\n
  • Ingestion points: The skill fetches metadata and search results from Google Scholar, PubMed, CrossRef, arXiv, and user-provided identifiers in extract_metadata.py.\n
  • Capability inventory: It executes local Python scripts that perform network requests and write formatted output to the file system.\n
  • Boundary markers: The instructions do not specify the use of delimiters or instructions to ignore embedded commands for the external data processed.\n
  • Sanitization: No content validation or sanitization of the retrieved metadata is described before its use in generation tasks.\n- [COMMAND_EXECUTION]: The skill utilizes several local Python scripts to perform search, extraction, and formatting operations, such as scripts/search_pubmed.py and scripts/format_bibtex.py.\n- [EXTERNAL_DOWNLOADS]: The skill requires the installation of well-known Python packages from official registries, including requests, biopython, scholarly, and selenium, to interact with academic APIs.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 05:39 AM