Security audit
Security checks for vulnerabilities and agentic risk
This content-rewriting skill is mostly coherent, but it requests broad command execution and includes out-of-scope network troubleshooting while only partially disclosing external LLM data handling.
Review this before installing if your agent grants skills shell access. Prefer local mode for sensitive or proprietary text, use LLM mode only when you are comfortable sending the content to the configured model provider, and consider removing or disabling exec/network troubleshooting unless you explicitly want the agent to debug connectivity.
The skill is for content rewriting, but its troubleshooting guidance instructs the agent to execute network diagnostics such as ping and inspect firewall/proxy settings. That expands behavior from text transformation into host and network interaction, increasing the chance of unnecessary command execution and environmental probing when handling user requests or failures.
The document claims the skill only rewrites content and does not validate facts, yet it also directs environment and connectivity checks outside that scope. This mismatch can cause an agent to perform broader operational actions than users expect, weakening least-privilege boundaries and making unsafe tool use more likely during error handling.
LLM mode sends user-provided content to an external model service, but the skill does not clearly warn users about that data transfer at the point where the mode is described. Sensitive or proprietary text could therefore be transmitted off-platform without informed consent, creating privacy, compliance, and data-handling risks.
