cancer-discovery

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill content is restricted to legitimate academic advisory information. No malicious code, obfuscation, or unauthorized network activities were detected.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest manuscript data and perform web searches, creating a surface for potential indirect prompt injection.
  • Ingestion points: Manuscript details and content provided by the user.
  • Boundary markers: Absent.
  • Capability inventory: The skill prompts the agent to perform external web searches for updated journal guidelines.
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 11:33 AM
Security Audit — agent-trust-hub — cancer-discovery