browser
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill interacts with external web content, creating a surface for potential injection attacks from malicious websites.
- Ingestion points: Data is retrieved from websites using
browser.snapshotandbrowser.scrapeinSKILL.md. - Boundary markers: The skill uses structured JSON and element handles (@eN) to prevent the agent from directly executing web content as instructions.
- Capability inventory: The skill can navigate to any URL and interact with elements, but requires explicit user confirmation for actions.
- Sanitization: Instructions explicitly forbid the use of arbitrary JavaScript or shell commands and recommend summarizing output to the user.
Audit Metadata