prioritize-accounts

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • Local Presentation Layer: The skill generates an HTML file in system temporary directories and can launch a local HTTP server to display account priorities. These features are used to provide an interactive user experience.
  • Sensitive Information Integration: It accesses and summarizes data from CRM, email, and meeting transcripts to provide sales context. The skill establishes CRM as the authoritative source of truth to ensure data reliability.
  • Content Security Measures: The instructions include a specific requirement to escape character sequences in JSON data to protect the generated view from potential script injection.
  • Data Processing Considerations: By ingesting content from external communication channels, the skill has an indirect prompt injection surface. It addresses this by providing clear rules against fabricating data and using multi-source verification.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 09:18 AM
Security Audit — agent-trust-hub — prioritize-accounts