Security audit
Security checks for vulnerabilities and agentic risk
This sleep-wellness skill is coherent and non-executable, with a language-localization safety note but no evidence of malicious or overbroad behavior.
Reasonable to install for general sleep-habit coaching. Users should treat it as wellness guidance only, seek medical help for persistent or severe sleep problems, and be aware that some crisis/disclaimer wording and examples are Chinese-localized rather than clearly language-adaptive.
The skill hard-codes crisis escalation and disclaimer content in Chinese without indicating that the response should match the user's language or offer alternatives. In a safety-critical moment, users who do not read Chinese may not understand self-harm escalation instructions, which can delay access to urgent help and undermine the model's duty of care.
The skill description includes substantial user-facing example content in Chinese, but the README does not state that the skill is Chinese-language only or offer users a language/locale choice. Under the policy rule, forcing or implicitly assuming a specific language without opt-in can be a natural-language policy issue.
No suspicious patterns detected.
