Security audit
Security checks for vulnerabilities and agentic risk
This is a Chinese dialogue interpretation skill with no code execution, persistence, credential use, or hidden data flows; its risks are mainly around overconfident interpersonal advice.
Install this if you want help interpreting Chinese conversation subtext, but treat outputs as subjective social readings rather than facts, especially for relationships, work disputes, legal issues, or major decisions. Provide only conversation text you are comfortable sharing with the agent.
The trigger list includes broad phrases like '她什么意思' and '帮我分析这段对话' that can appear in ordinary conversation and may cause unintended invocation. Over-broad routing can expose user content to the wrong skill, leading to privacy issues, misclassification, or use of a manipulative interpretation mode in contexts where the user did not explicitly request it.
Early instructions say that when information is insufficient, the skill should stop and ask follow-up questions rather than speculate, implying a high-confidence interpretation standard. However, the mandatory closing disclaimer says '以上全凭你发的这几句话脑补' and '准不准还得你自己拍板,' which characterizes every output as conjectural even when the earlier workflow says sufficient context has been established. This is an intent-level contradiction within the documentation.
The skill instructs the agent to produce a definitive '确定结论' and avoid '可能也许' wording in the structured output. But later it states that interpretation is '概率性洞察, 不是事实断定' and even mandates a disclaimer saying the result is based on limited text and may be inaccurate. These instructions conflict on the intended certainty level of the skill's output.
The entire file is written as Chinese-only guidance and explicitly frames interpretation through a '中式语境' lens, which imposes a specific language/cultural locale without any opt-in or alternative-language support. Under the policy, locale-specific behavior should either offer user choice or be clearly documented as a justified regional constraint.
The description and title define the skill exclusively around interpreting Chinese dialogue and expressions, but do not explicitly frame this as a user-selected language/locale specialization or offer alternatives. This can be a language-policy issue when a skill imposes a specific language context without clear opt-in or justification.
No suspicious patterns detected.
