Security audit
Security checks for vulnerabilities and agentic risk
This is a simple communication-style skill that is broadly coherent, with a usability caution because brief-mode triggers are broad and the style persists until turned off.
Install this only if you want a terse assistant style that may stay active after phrases like "be brief" or "less tokens." Use "stop caveman" or "normal mode" to turn it off, and be cautious in contexts where full wording matters.
The trigger phrases include broad, common-language requests like "be brief" and "less tokens," which can activate the mode unintentionally during normal conversation. Because the mode then persists across future responses, accidental activation can degrade clarity and safety-sensitive communication, especially if users do not realize a persistent behavior change occurred.
The skill forces a persistent communication mode once triggered and states it remains active indefinitely until the user explicitly disables it. This creates a communication-policy risk because the assistant may continue using compressed, fragmentary language in later contexts where clarity, nuance, or complete safety messaging is important.
No suspicious patterns detected.
