ask-user-question

Pass

Audited by Gen Agent Trust Hub on Apr 3, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No malicious instructions or attempts to bypass system constraints were found. The instructions provide functional guidance for utilizing the UI interaction tool within its intended context.\n- [DATA_EXFILTRATION]: No access to sensitive files or exfiltration of user data was detected. Network communication is limited to a local endpoint (localhost:9227) used to bridge communication between the agent and the local UI interface.\n- [REMOTE_CODE_EXECUTION]: The skill does not download or execute remote scripts. It uses standard dependencies and does not include any dynamic code execution patterns.\n- [COMMAND_EXECUTION]: No dangerous shell commands or attempts at privilege escalation were found. The skill operates within the restricted environment of the Model Context Protocol SDK.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 3, 2026, 01:40 AM