oneshot-browser

Pass

Audited by Gen Agent Trust Hub on Nov 1, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill enables the agent to navigate and interact with arbitrary websites, which are untrusted sources. Web content can be crafted to include hidden instructions that attempt to override the agent's goals or exfiltrate data processed during the session.
  • Ingestion points: The agent.browser function processes natural language tasks by interacting with web content from specified URLs.
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard instructions found within the retrieved web data.
  • Capability inventory: The skill supports navigation, form filling, structured data extraction, file downloads, and the use of persistent session profiles with stored secrets.
  • Sanitization: No sanitization or validation logic is presented for the content ingested from the browser sessions.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 1, 2026, 06:34 PM