openai-screenshot
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHCOMMAND_EXECUTIONDATA_EXFILTRATIONREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION] (HIGH): The skill metadata explicitly instructs the agent to seek escalated permissions if it encounters sandbox restrictions, promoting privilege escalation.
- [DATA_EXFILTRATION] (HIGH): The skill provides tools for system-wide screenshots, creating a high risk of sensitive data exposure.
- [REMOTE_CODE_EXECUTION] (MEDIUM): The PowerShell script uses dynamic C# compilation (Add-Type) to access system APIs, which is a common vector for code injection.
Recommendations
- AI detected serious security threats
Audit Metadata