Security audit
Security checks for vulnerabilities and agentic risk
This skill is a coherent local workflow for reviewing Chinese customer-service chat logs and producing masked quality-control reports.
Install this if you need Chinese customer-service conversation QC. Because it handles customer chat logs, review the generated report before using it for employee assessment and avoid sending private conversation data through the optional public feedback channels.
The version note explicitly mentions a clarified '中文会话场景声明', which indicates the skill is constrained to Chinese-language conversations. The README does not present this as a user choice or opt-in, and no justification is given here for enforcing a specific language/locale.
The skill’s applicability text is unusually broad and explicitly claims it should be used even when the user does not mention '质检'. That can cause over-triggering on generic chat-review or evaluation tasks, leading the agent to process sensitive conversation exports under an incorrect workflow and with assumptions the user did not explicitly request. In a privacy-heavy skill that handles customer chat logs, overly broad routing increases the chance of unintended data exposure, wrong task execution, or inappropriate scoring outputs.
The entire dataset, including headers and conversation content, is presented only in Chinese with no indication that language choice is optional or contextually justified. Under the stated policy, a skill artifact that enforces a single language without user opt-in can constitute a natural-language policy violation.
This markdown file is subject to natural-language policy checks. The title and all rule content are presented exclusively in Chinese, which can constitute a language/locale policy violation when no user language choice, opt-in, or region-specific justification is provided.
The entire guide is written in Chinese and presents its instructions as the default operating language, with no indication that users may choose another language or locale. Under the policy rule for natural-language violations, a skill that imposes a specific language without opt-in should be flagged unless the locale constraint is explicitly justified.
The README suggests activating the skill by saying “帮我质检这批客服会话”, but it does not clearly define whether this is the exact trigger, one example, or what similar phrases should or should not invoke the skill. Without explicit scope or negative examples, users and routing logic may interpret a broad quality-check request too loosely.
The privacy section states that data is processed locally and 'this skill does not upload any data' (L33). Later, the README invites users to use GitHub Issues and email for feedback (L44-L45), which are external destinations; while not code behavior, this creates a mild documentation-level contradiction about data locality and external communication.
The manifest description ends with a statement that the rule package is designed for Chinese conversations. While later text notes that scoring may be limited for other languages, the skill description itself presents a locale-specific constraint without an explicit opt-in or alternative path.
This markdown file uses only Chinese for headings, tables, and warnings, which can amount to a language/locale policy issue when no user opt-in or documented regional justification is provided. The stated policy asks to flag skills that force a specific language without offering a choice.
No suspicious patterns detected.
