Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed sales budget research aid that uses public-source company research and local calculations, with no persistence or credential handling, though its account-intelligence outputs should be treated as speculative.
Install only if you want a Chinese-language sales research workflow that searches public company information and generates inferred budget and stakeholder-role guidance. Treat decision-maker and budget outputs as estimates, verify sources before acting, and avoid using it to collect or infer private personal information.
The skill name and description are entirely in Chinese, which implies a fixed language experience. The file does not indicate that users may choose another language or that the skill is intentionally limited to a Chinese-speaking region or audience.
The skill instructs the agent to gather external data, infer budgets, map suppliers, and identify key decision-makers, but it does not warn users that some outputs are inferred and may involve sensitive business or personal-professional information. Without disclosure and guardrails, users may unknowingly generate speculative intelligence profiles that create privacy, accuracy, and misuse risks.
The trigger section uses broad, natural sales-language prompts such as whether a customer is worth following up or spends on digitalization, which can cause the skill to activate in contexts broader than narrowly scoped budget analysis. That increases the chance of unintended collection, inference, and presentation of sensitive company and decision-maker intelligence without a clear user intent boundary.
SQP-3 applies to all file types and covers language or locale policy violations. This markdown file presents all test instructions and outputs exclusively in Chinese, with no opt-in, alternative language, or justification for a Chinese-only scope.
Manifest 将技能定位为企业数字化预算情报工具，输出范围限于预算估算、构成分析、采购周期和销售建议。但测试用例明确覆盖“关键决策人分析”“行业趋势分析”“机会点生成”以及包含这些内容的“完整报告生成”，说明实际技能能力明显超出描述范围。
This code includes natural-language descriptions and user-facing output primarily in Chinese, such as the module docstring and CLI text, but does not offer the user a language/locale option. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is clearly documented and justified.
This code goes beyond generic budget estimation and produces target-account intelligence such as likely decision-maker roles, vendor landscape, procurement timing, and account-specific opportunity guidance. In a sales/CSM skill, that materially increases the sensitivity of the output because it can be used to profile organizations and optimize influence over procurement stakeholders, even though it does not extract private data or perform exploitation directly.
The calculate_budget_composition docstring states it returns a structure refined to categories and suppliers, and the function indeed returns nested objects with ratio and vendors. However, the test CLI later prints each entry as v*100, implying the value is a bare numeric percentage; this contradicts the documented and actual return shape and would fail at runtime.
This manifest-style JSON uses Chinese-only natural-language descriptions for tools and parameters throughout the schema. Under the policy for natural-language violations, forcing a specific language without user opt-in or a documented region-specific justification is a locale-policy issue.
No suspicious patterns detected.
