Security audit
Security checks for vulnerabilities and agentic risk
This skill is a simple Chinese e-commerce reply-template helper with some quality issues, but no hidden execution, credential access, persistence, or destructive behavior.
Use this skill for Chinese-language e-commerce after-sales replies and review the generated message before sending. Remove or hide the internal-notes section from buyer-facing communications, and localize the templates and closing for other languages or marketplaces.
The description says to use the skill for return, exchange, refund inquiries and broadly for "any after-sales consultation," which is wider than the specific refund/return scope implied by the skill name. It does not provide explicit trigger boundaries or negative examples, increasing the chance of unintended invocation for general support chats.
The skill contains conflicting instructions: the workflow says to output only one customer-facing message with no meta-commentary, while the output format requires an additional internal-notes section. In practice, this can cause an agent to leak internal classifications, policy checks, or template details to the buyer, creating information disclosure and reliability issues in customer-facing communications.
The instruction to always end with the Chinese phrase "如有其他问题，随时联系我们 😊" forces a specific language regardless of the user's language or locale. This can violate language/locale policy when the skill is used in multilingual contexts without user opt-in or a documented region-specific justification.
This markdown file contains customer-facing reply templates exclusively in Chinese across all issue types. Under the language/locale policy, forcing a specific language without user opt-in or a documented region-specific justification is a natural-language policy violation.
No suspicious patterns detected.
