Security audit
Security checks for vulnerabilities and agentic risk
This is a prompt-only Chinese counterfactual life-story skill with no code, external access, persistence, or hidden execution behavior.
Install only if you want a Chinese-language counterfactual life-story prompt that may infer personality and background from short inputs. Be cautious with very personal prompts, and correct the assistant if its assumptions about your culture, age, location, or past choices are wrong.
The trigger list includes very common phrases such as '假如' and '如果当初', which can appear in ordinary conversation unrelated to this skill. That makes accidental activation likely, causing the agent to switch into speculative life-simulation behavior without clear user intent and potentially generating unwanted inferences about the user's personality or past choices.
The instruction to 'directly start' from a single short phrase and 'not ask follow-up questions' removes an important safety boundary for confirming user intent. In practice, this increases the chance of misfires and pushes the model to immediately generate sensitive, personalized speculation from minimal context.
The skill assumes users are 'mainly Chinese' and hard-codes cultural and locale-specific background into its reasoning. This can cause inaccurate personalization, stereotyping, and inappropriate inferences when the user is not Chinese or does not want that framing, especially since the skill is designed to infer personality and life trajectory.
No suspicious patterns detected.
