headline-generator

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a pure text generation tool that transforms user-provided facts into headline candidates. It does not utilize any tools, execute shell commands, or perform network operations.
  • [SAFE]: External URLs within the documentation point to well-known academic and scientific publishing domains (e.g., Nature, PMC, University of Chicago, University of Toronto) and are used solely for research provenance.
  • [SAFE]: Analyzed the indirect prompt injection surface. Untrusted data enters the agent via user-provided story facts (Ingestion Point). Boundary markers and input sanitization are not present. However, the skill has no active capabilities (Capability Inventory: None), meaning any injected instructions would not have an environment in which to execute.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 01:09 AM
Security Audit — agent-trust-hub — headline-generator