verify-citation

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill retrieves academic citation metadata and abstracts from the official APIs of CrossRef (api.crossref.org) and OpenAlex (api.openalex.org). These are well-known and trusted academic data providers. The interaction is limited to standard GET requests for citation data, consistent with the skill's stated purpose.
  • [COMMAND_EXECUTION]: The skill is executed as a standalone Python script and does not involve any subprocess calls, dynamic command execution, or shell-related operations, maintaining a narrow and safe operational scope.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests paper abstracts from external APIs and presents them to the agent.
  • Ingestion points: Local BibTeX files and JSON responses from academic APIs.
  • Boundary markers: Abstract content is presented within clearly labeled sections in the script's output report.
  • Capability inventory: The skill can update or create a project summary CSV file (references.csv).
  • Sanitization: Input parameters for API requests are properly URL-encoded using standard library functions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 01:13 PM
Security Audit — agent-trust-hub — verify-citation