The Agent Skills Directory

[PROMPT_INJECTION]: The skill employs strong persistence instructions, such as 'ACTIVE EVERY RESPONSE' and 'No revert after many turns', to ensure the agent maintains the 'caveman' persona across long conversations. While this overrides the default AI conversational style, it is focused on formatting and efficiency rather than bypassing safety protocols.
[SAFE]: The skill incorporates an 'Auto-Clarity' rule that serves as a security safeguard. It explicitly instructs the agent to drop the compressed style and return to normal prose for security warnings, confirmations of irreversible actions, and complex multi-step sequences where ambiguity could lead to user error. This practice ensures that critical safety and technical information remains clear and unambiguous.
[SAFE]: No evidence of data exfiltration, credential harvesting, obfuscation, or remote code execution was found. The skill operates entirely within the scope of character manipulation and prose style adjustment.

caveman