Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only mindfulness guidance skill with clear wellness boundaries and no executable, persistent, or data-access behavior.
This skill appears safe to install for conversational mindfulness support. Be aware it is not medical care, it uses mostly Chinese examples and China-specific crisis resources, and users in other regions should rely on local emergency or crisis services when safety is at risk.
The README presents all user-facing skill behavior and examples exclusively in Chinese. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly justified, which is not present here.
The file mixes English headings with overwhelmingly Chinese user examples, response examples, disclaimer text, and crisis escalation language, and it instructs expected outputs in Chinese without stating that language selection should follow user preference. This creates a natural-language policy concern because the skill appears to default to a specific language/locale rather than offering opt-in or choice.
No suspicious patterns detected.
