citation-management

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's instructions are focused on data validation, normalization, and verification of academic citations against reputable sources. No patterns of prompt injection, data exfiltration, or unauthorized command execution were detected.
  • [SAFE]: The skill processes user-provided BibTeX data, which is an attack surface for indirect prompt injection. However, the instructions provide strong guidance on parsing and verifying identifiers (DOIs, PMIDs) against authoritative metadata sources (Crossref, PubMed, etc.) rather than trusting the input data or search snippets.
  • [SAFE]: No external code dependencies or remote code execution patterns were found. The skill describes a logical workflow for an agent to follow using its existing toolset.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 01:19 PM
Security Audit — agent-trust-hub — citation-management