creating-replay-vision-scanners
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes session recordings and observation results, which are external data sources.\n- Ingestion points: The agent ingests data through tools like
vision-scanners-observations-listandvision-observations-listinSKILL.md.\n- Boundary markers: There are no explicit instructions for the agent to treat processed recording data or observation text as untrusted or to use specific delimiters.\n- Capability inventory: The skill provides the agent with the ability to create and modify scanners (vision-scanners-create,vision-scanners-update), create alerts (vision-alerts-create), and generate dashboards (vision-scanners-dashboard-create).\n- Sanitization: No specific sanitization or filtering logic is mentioned for the content of observations before they are presented to the user or used to trigger alerts.
Audit Metadata