slide-editing

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides standard formatting and styling guidelines for HTML-based presentation slides, using common CSS properties and structures.- [SAFE]: The skill instructs the agent to use designated tools like view-screen and update-slide for slide manipulation, explicitly advising against raw API calls.- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface. Ingestion points: Untrusted HTML data is read from the 'active deck' via the view-screen tool. Boundary markers: None are specified to separate data from instructions. Capability inventory: The skill uses the update-slide action to modify database content. Sanitization: No validation or filtering of external HTML is described. This surface is inherent to the slide-editing use case and does not involve high-risk capabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 04:51 AM
Security Audit — agent-trust-hub — slide-editing