audit-website

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted data fetched from external websites, which is a vector for indirect prompt injection attacks where a malicious website could attempt to influence the agent's behavior. \n
  • Ingestion points: Website content and DOM snapshots are fetched from arbitrary URLs via web_fetch and browser tools as defined in SKILL.md. \n
  • Boundary markers: The instructions do not define specific delimiters or isolation warnings to distinguish fetched content from the agent's logic. \n
  • Capability inventory: The skill utilizes web_fetch, browser (open/snapshot), and the squirrel CLI tool. \n
  • Sanitization: No explicit sanitization or filtering of external content is mentioned. \n- [COMMAND_EXECUTION]: The skill references and utilizes an external CLI tool named squirrel for performing audits, which involves executing code on the host system with parameters derived from user-provided URLs.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 05:29 PM
Security Audit — agent-trust-hub — audit-website