normies

Pass

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill employs persona-based instructions to guide the agent in simulating different user perspectives. These role-playing directives are essential for the skill's purpose and do not conflict with AI safety protocols.
  • [PROMPT_INJECTION]: As the skill visits external URLs provided by the user, it is exposed to third-party content which represents a surface for indirect prompt injection. This is a functional requirement for its task.
  • Ingestion points: Site content at the user-provided URL accessed via browser tools.
  • Boundary markers: The agent is explicitly told to disregard source code, focusing only on the rendered interface.
  • Capability inventory: Limited to browser navigation, screenshots, and textual reporting.
  • Sanitization: Content is processed as rendered by the browser; no additional sanitization is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 12, 2026, 07:23 AM