Security audit
Security checks for vulnerabilities and agentic risk
This security skill is purpose-aligned, but it advertises automatic process killing and recurring outbound security reports without clear approval, scoping, or disable controls.
Review this skill before installing on production or personal machines. Use it only where you are comfortable with local security inspection, and require explicit approval before killing processes, creating cron jobs, sending reports, or scanning secrets. Configure any report destination deliberately and disable recurring monitoring when it is no longer needed.
The skill advertises 'auto-kills resource drains' and zombie process cleanup without any warning, approval gate, or safety constraints. Automatic process termination on a host can kill legitimate services, disrupt workloads, or be triggered inappropriately by an overbroad interpretation of 'suspicious' or 'resource drain' behavior.
The skill describes recurring automated security audits with findings sent to a preferred channel, but does not warn users about persistent monitoring, scheduled execution, or outbound transmission of potentially sensitive host-security data. This can create privacy, data-leakage, and surprise-execution risks, especially on personal machines or production systems.
The invocation examples are very broad natural-language phrases that could match ordinary user requests without clearly signaling that this skill should take over. Because the skill advertises sensitive security and system-management behaviors, ambiguous activation increases the chance of unintended execution of audits or system-altering workflows.
No suspicious patterns detected.
