Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed request-analysis workflow with usability concerns from broad activation and Chinese-only templates, but no evidence of code execution, persistence, data access, or exfiltration.
Install this only if you want the agent to frequently pause for structured requirement analysis. Users who prefer concise answers or non-Chinese output should adjust the skill to mirror the user's language and narrow activation to complex or ambiguous tasks.
The activation description is extremely broad, triggering on ordinary user tasks, questions, link sharing, and vague phrases. This can cause the skill to intercept a large share of normal conversations and impose its workflow even when unnecessary, increasing the chance of prompt-routing abuse, user friction, and unintended control over agent behavior.
The skill content is written to enforce a Chinese-language workflow and output format without any indication that the user's language preference should be respected. In practice this can override user expectations, reduce usability, and create confusion or unsafe misunderstandings if the surrounding interaction is in another language.
The skill content is entirely written in Chinese and includes fixed Chinese response examples and stylistic directives, which can push the agent to respond in Chinese regardless of the user's preferred language. This creates a usability and safety issue because it can degrade clarity, prevent informed user confirmation, and interfere with higher-priority system or user language preferences.
