citation-management

Pass

Audited by Gen Agent Trust Hub on May 28, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a set of academic research tools that follow security best practices. All external interactions are directed towards well-known and trusted academic APIs (CrossRef, PubMed, arXiv) and a reputable AI provider (OpenRouter).
  • [COMMAND_EXECUTION]: The use of subprocess.run in scripts/generate_schematic.py follows best practices by using a list of arguments rather than a shell string, which prevents shell injection vulnerabilities. The script is used to modularize the execution of internal Python utilities.
  • [EXTERNAL_DOWNLOADS]: The skill connects to external APIs to retrieve citation data and generate images. These downloads are integral to the stated purpose of the skill and target established, reputable academic domains.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied descriptions to generate scientific diagrams. While this is an ingestion point for untrusted data, the capabilities are appropriately scoped, and the risk is assessed as low and inherent to the functionality.
  • [CREDENTIALS_SAFE]: Secret management is handled correctly via environment variables (e.g., OPENROUTER_API_KEY, NCBI_API_KEY), following standard security practices to avoid hardcoded credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
May 28, 2026, 03:41 PM
Security Audit — agent-trust-hub — citation-management