prioritize-accounts
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- Local UI Generation: The skill writes a temporary HTML file to provide an interactive account action view. This is a functional requirement and includes JSON escaping to prevent injection attacks within the generated dashboard.
- Indirect Prompt Injection Surface: The skill processes information from external sources like CRM, calendars, and messaging platforms (Ingestion points). The potential for data-driven instructions is mitigated by requirements for source-grounded evidence (Boundary), the use of structured output categories (Capability), and implementation of JSON escaping and HTML encoding (Sanitization) to ensure data integrity.
- Local Hosting Fallback: To ensure dashboard accessibility, the skill may initiate a local HTTP server if direct file linking is unavailable. This process is confined to the local machine and serves only to render the prioritization results for the user.
Audit Metadata