caveman
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No malicious instruction overrides or safety bypasses detected. The skill provides behavioral instructions to alter the agent's communication style (brevity) but explicitly includes a mechanism to drop this style for security warnings and irreversible actions.
- [DATA_EXFILTRATION]: No network operations, sensitive file access, or credential exposure detected.
- [REMOTE_CODE_EXECUTION]: No external package dependencies or remote script execution patterns found.
- [COMMAND_EXECUTION]: No shell commands, subprocess spawning, or privilege escalation attempts detected.
- [OBFUSCATION]: The content is clear and readable with no hidden characters, encoding, or deceptive formatting.
Audit Metadata