Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only advisory skill that gives structured business-planning and decision-review responses in a Chinese courtly style, with no executable behavior or data-access mechanism found.
Install this if you want structured business and planning advice in a Chinese courtly tone. Before relying on it for pricing, hiring, contracts, or major business decisions, ask for assumptions, evidence, and validation steps, and explicitly request a neutral tone or another language if the default honorific style is not desired.
This markdown file contains user-facing natural-language instructions and examples exclusively in Chinese, including the recommended invocation phrases. Under the policy rule for language/locale, forcing a specific language without user opt-in can be a violation when no alternative or justification is provided.
The entire skill file is written in Chinese and presents its positioning, messaging, and user-facing content exclusively in that language. There is no indication that users may choose another language or that the Chinese-only constraint is required for a documented region-specific purpose, which can violate language/locale policy expectations.
The recommended first utterance '帮我做一次太子分案' is presented as a direct trigger, but the document does not define clear scope, exclusion conditions, or negative examples for when this invocation should or should not activate. Because the phrase is framed as a general-purpose starter for many situations, it risks ambiguous or unintended invocation behavior.
The '10 starter phrases' section lists many natural-language phrasings as suitable openers, but it does not indicate whether these are exact triggers, examples only, or how much variation is allowed. This ambiguity can cause overly broad matching against everyday conversational requests, especially for phrases like asking for help judging priorities or reviewing a matter from a given angle.
The skill sets a default honorific and forbids directly addressing the user as “you” unless the user explicitly requests otherwise, which constrains output style without prior user opt-in. This is not a code-execution or data-exfiltration issue, but it can override user preference, reduce accessibility, and create unwanted coercive tone or social pressure in interactions.
The skill hard-codes honorific forms of address and explicitly discourages ordinary second-person pronouns unless the user opts in. While not a classic security flaw, it creates a behavioral constraint that can override user preference and reduce user autonomy, which is a legitimate policy/control concern in agent skills. In this context, the impact is limited because the instruction is stylistic rather than enabling code execution, data exfiltration, or unsafe actions.
The trigger conditions include broad phrases like '用户问题模糊' and '用户同时抛出多个事项', which overlap with many ordinary conversations and do not clearly define when this role should activate versus not activate. The file provides no explicit trigger phrases, constraints, or negative examples to limit invocation scope.
Conditions such as '用户有模糊方向，需要成形' and '用户需要提案、方案、框架、表达' are generic and could match a very large share of user requests. Without explicit boundaries or examples, the role may be invoked too broadly.
Phrases like '用户在做重大判断' and especially '用户可能自我感动' rely on subjective interpretation and do not provide a clear, testable threshold for activation. This ambiguity can cause inconsistent or unintended routing.
Triggers such as '用户已经基本想清楚' and '用户需要行动顺序' are common across many advisory conversations and are not limited by context or exclusions. The absence of negative examples makes overlap with other roles likely.
The text states a default form of address for the user as specific Chinese honorifics, which imposes a language/locale style choice by default rather than offering it as an option. The policy allows this only when the skill offers user choice or clearly justifies the locale constraint, which is not present here.
The statement '语言可以贵，但结构必须清楚' imposes a particular stylistic/language preference as a general standard rather than offering it as an option based on user preference. This may conflict with policy against forcing a specific language or locale style without opt-in.
No suspicious patterns detected.
