e2etest
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill utilizes agent-browser and Playwright to navigate and interact with external websites. This interaction exposes the agent to untrusted data from the web, which could contain malicious instructions intended to subvert the agent's logic during testing. * Ingestion points: External web content accessed via browser tools in SKILL.md. * Boundary markers: Not specified; the instructions do not provide guidance on distinguishing between test commands and data from the page. * Capability inventory: Execution of CLI tools (agent-browser), element interaction, and screenshot capture. * Sanitization: No validation or sanitization of external content is described.
- [EXTERNAL_DOWNLOADS]: The skill references documentation and releases for the agent-browser tool from the vercel-labs GitHub repository and associated official domains. These are well-known and trusted sources for development resources.
Audit Metadata