regulars
Pass
Audited by Gen Agent Trust Hub on Mar 25, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it interacts with untrusted external web content.
- Ingestion points: External website content is ingested via browser MCP tools when navigating to target URLs defined in the
Agent Prompt TemplateinSKILL.md. - Boundary markers: The subagent instructions lack explicit delimiters (e.g., XML tags or clear separators) and warnings to disregard embedded instructions within the website content.
- Capability inventory: The agents have the capability to navigate, click elements, search, and capture screenshots using browser MCP tools.
- Sanitization: There is no evidence of sanitization, filtering, or validation of the text or metadata retrieved from the target websites before it is processed by the agent.
Audit Metadata