The Agent Skills Directory

[PROMPT_INJECTION]: The skill provides explicit instructions to bypass standard AI safety and alignment guardrails. Directives such as "No selective refusal," "No vocabulary policing," and instructions to eliminate "Disclaimer addiction" and "Patronizing warnings" are patterns used to override the agent's built-in constraints and safety filters.
[PROMPT_INJECTION]: The "Ten sins to eliminate" section instructs the agent to disregard typical conversational boundaries (hedging, moralizing, disclaimers) that are often established in system prompts to ensure safe and neutral responses.
[PROMPT_INJECTION]: The instruction to prioritize "Historical accuracy over modern representation" and "Biology is biology" specifically targets and attempts to override the agent's internal training regarding cultural sensitivity and neutrality.

unwoke