Security audit
Security checks for vulnerabilities and agentic risk
This skill is a local Chinese-language customer handoff draft generator with no evidence of hidden networking, persistence, or destructive behavior.
Use this for Chinese-language sales-to-delivery handoff drafts, and review the output before sharing it. Only point the script at files you intentionally want summarized, especially if they contain customer or personal data; the packaged behavior is local, but the script accepts arbitrary input and output paths.
声明描述的是一个面向客户交接的业务整理工具，核心应聚焦销售到交付的信息梳理、承诺风险识别、实施前提提炼。代码却没有任何与客户、销售承诺、交付前提、CS 交接语义相关的专门逻辑；相反，它实现的是一个通用报告生成/审计器，并支持目录扫描、CSV 字段概览、正则风险模式检测以及 Skill 结构合规检查。这些能力与声明用途存在明显偏离，且主用途更像本地内容/代码审计工具而不是 sales-to-cs handoff 助手。虽然 structured_brief 模式理论上可被 spec 配置成某种交接模板，但仅从代码看，其实际能力范围显著更广且主要是通用审计，因此应判定为描述与行为不匹配。
The pattern scanning feature searches arbitrary files and directories for secrets, private URLs, and shell-execution indicators, which is a security-auditing function unrelated to account handoff generation. In this skill context, that creates an unjustified capability to inspect local repositories and sensitive text sources, increasing the chance of unauthorized discovery of credentials or internal infrastructure details.
The dispatcher exposes multiple generic audit modes such as directory, CSV, pattern, and skill auditing that go well beyond the stated purpose of building customer handoff packages. This capability expansion materially increases the skill’s power to inspect arbitrary local content, enabling misuse for repository reconnaissance and unrelated analysis of user-provided paths.
整个 README 以中文呈现，且未说明该技能是否仅面向中文用户，或是否支持按用户偏好切换语言。按规则，若技能在语言或区域设置上施加强约束而未提供选择或合理说明，可构成自然语言政策违规。
The trigger examples are short, generic phrases such as '把销售资料整理成客户交接包' and '找出承诺风险', which can overlap with ordinary user requests in adjacent business contexts. In an agent routing environment, overly broad triggers can cause the skill to activate unintentionally and process sensitive sales or customer materials without the user explicitly intending to invoke this workflow.
This markdown file presents all user-facing audit and status information exclusively in Chinese. Under the policy rule for natural-language violations, forcing a specific language without user choice can be a locale/language policy issue when no opt-in or justification is provided.
The skill advertises executable capability via python3 and references reading templates/specs and writing output, but it does not declare explicit tool scope such as allowed tools or permissions. That creates an authorization gap: a host agent may permit broader shell or filesystem access than users expect, increasing the chance of unintended file access or command execution.
The manifest description and primary skill instructions are written entirely in Chinese, and the skill does not state that language can be adapted to the user's preference. This creates a natural-language policy concern because it implicitly enforces a specific language/locale without opt-in or an explicit region-specific justification.
The file is entirely written in Chinese and presents the skill input structure and requirements only in that language, with no indication that users may choose another language or that the skill is intentionally limited to a Chinese-speaking context. This can violate language/locale policy when a skill implicitly forces one language without opt-in or documented justification.
This markdown example output forces a specific language/locale across the entire skill content, and there is no indication that the user can choose another language. The policy explicitly calls for flagging language or locale constraints when they are imposed without user opt-in or documented justification.
The user-facing title, summary, input hint, sections, examples, and risk text are all specified only in Chinese, with no indication that the skill supports other languages or lets the user choose a locale. This can violate language/locale policy when a skill implicitly constrains interaction to a specific language without documented opt-in or region-specific justification.
The entire template, including the title and section instructions, is written only in Chinese, which imposes a specific language on users. The file does not offer a language choice or explain that the skill is intentionally limited to a Chinese-speaking or region-specific context.
The generated report strings in make_structured_report are written entirely in Chinese, and similar hard-coded Chinese output appears elsewhere in the script. This forces a specific language on users without opt-in or justification, which matches the language/locale policy violation criteria.
directory_report and csv_report emit headings and explanatory text only in Chinese, again imposing a locale choice across multiple execution modes. There is no user-selectable locale setting or documented reason for this restriction in the file.
pattern_report, parse_frontmatter error messages, and skill_audit all generate user-facing text in Chinese only. Because this behavior spans the script and offers no language choice, it constitutes a natural-language policy issue rather than an isolated string choice.
The skill_audit and frontmatter validation logic can inspect arbitrary skill directories and enumerate expected files, which is unrelated to customer handoff packaging. While not directly code-executing, it enables filesystem inspection and metadata extraction outside the declared business purpose, broadening the attack surface and facilitating reconnaissance on local project structure.
This markdown file presents all instructions and test expectations exclusively in Chinese. Under the policy rule for natural-language violations, forcing a specific language without user opt-in or a documented regional justification is a reportable issue.
No suspicious patterns detected.
