Security audit
Security checks for vulnerabilities and agentic risk
This is a Chinese-language structured-thinking prompt pack with no code execution, network access, credential use, or persistence; its main caveat is that broad trigger words may activate it more often than intended.
This skill appears safe to install as a prompt-only thinking aid. Before using it, expect Chinese-language flows and confirm the selected module when the request is broad, especially for generic analysis, prioritization, or decision questions. Treat its strategic, decision, and productivity outputs as structured guidance, not as a substitute for professional advice in high-stakes situations.
The trigger list for 根因分析 includes very generic words such as “为什么”, “分析”, and “怎么回事”, which commonly appear in ordinary conversation outside the intended domain. Because the file presents these as activation keywords without clear exclusion conditions, the skill could be invoked in contexts where the user is not actually asking for this framework.
The keyword list includes broad terms like “市场”, “规划”, and “定位”, which can appear in many user requests that are not specifically seeking a SWOT or five-forces analysis. The routing description does not provide negative examples or tighter scope constraints to prevent accidental matching.
Terms such as “选择”, “评估”, “哪个好”, and “怎么选” are common in everyday dialogue and are not sufficiently specific on their own to indicate this decision-analysis skill. The document lacks clear limits or negative examples defining when these phrases should not activate the module.
This trigger list contains many generic expressions such as “思考”, “分析”, “不清楚”, “困惑”, and “没头绪”, which are common in normal user conversation and not unique to this module. Without tighter boundaries or exclusion conditions, the router may over-select this skill for vague problem statements.
Keywords like “重点”, “关键”, “效率”, and “先做什么” are widely used across many contexts and are not uniquely indicative of an 80/20 or energy-allocation analysis. The file does not define boundaries or exclusions to distinguish everyday phrasing from intended activation cases.
The entire skill flow, prompts, and output templates are written as mandatory Chinese-language interaction, with no indication that the user can choose another language or that the module is restricted to a Chinese-speaking context. This creates a natural-language locale policy concern because it implicitly enforces one language without opt-in or documented justification.
The skill is written entirely in Chinese and prescribes Chinese-language prompts without providing any language selection or fallback behavior. This can cause user confusion, exclusion, or misuse when routed to users who do not understand Chinese, reducing informed consent and reliability of the interaction rather than creating a direct code-execution or data-exfiltration risk.
Earlier sections repeatedly state that the module must only present quantified data and let the user decide, explicitly prohibiting subjective recommendations. However, the final output template says '保持或加大投入' and '建议改法', which are direct recommendations rather than neutral data presentation, contradicting the documented boundary and core principle.
The document says all judgments must be based on user-provided numbers and that the module is a quantitative 80/20 tool, not subjective advice. Yet the exception handling instructs that when the user does not want itemized quantification, '其他默认低效', which assigns a low-efficiency conclusion without collecting the promised data.
模块A的关键词包括“为什么”“原因”“分析”等常见表达，这些词在普通问答中极其常见，若缺少边界约束会导致根因分析模块被频繁误路由。误触发虽不直接造成代码执行或数据泄露，但会系统性影响技能行为，使输出偏离用户真实意图。
模块D包含大量高频、语义宽泛的触发词，如“分析”“讨论”“不懂”“困惑”等，容易在与多种任务无关或仅弱相关的对话中被误触发。对路由层而言，这会造成过度捕获用户请求，降低模块选择准确性，并可能把用户导向不适合的分析流程。
This markdown file contains all interaction scripts and output templates exclusively in Chinese, including the exact user-facing prompts the skill should use. Because it does not offer a language/locale choice or document that the skill is intentionally region-specific, it may violate a policy requiring user language preference or opt-in.
“语义理解兜底”没有说明适用边界、证据标准或失败处理方式，意味着模型可在弱信号下自由解释用户意图，增加不可预测路由的概率。该问题主要影响一致性和可控性，可能放大前述宽泛关键词带来的误判。
