hero-infinite-marquee-stack

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides boilerplate code for a layered marquee hero section using standard browser APIs like requestAnimationFrame, ResizeObserver, and CSS transforms. No malicious patterns, such as obfuscation, credential harvesting, or unauthorized network access, were detected.
  • [INDIRECT_PROMPT_INJECTION]: The implementation performs DOM manipulation by reading innerHTML from elements to duplicate content for the marquee effect. While this creates a surface where content from the DOM is processed, the skill lacks any capabilities to exfiltrate data or execute system commands.
  • Ingestion points: The duplicateContent function and the reference implementation read from the .marquee-track container using innerHTML.
  • Boundary markers: None present.
  • Capability inventory: The skill is limited to UI manipulation; it contains no network access, file system access, or command execution capabilities.
  • Sanitization: None; content is cloned directly to maintain visual fidelity of the typographic layout.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 05:48 PM
Security Audit — agent-trust-hub — hero-infinite-marquee-stack