Security audit
Security checks for vulnerabilities and agentic risk
This is a prompt-only Chinese procurement review skill that analyzes user-provided business-registration relationships and does not install code, persist data, or autonomously call external services.
Before installing, users should understand that this skill may process commercially sensitive bidder information and personal contact details if supplied. Use redacted or public business-registration data where possible, and treat its conclusions as review assistance only, not as a final legal or procurement decision.
The entire skill description, triggers, input/output guidance, and usage instructions are written only in Chinese, and there is no statement that the skill supports other languages or requires Chinese by design. Under the policy, forcing a specific language without user opt-in is a natural-language policy issue unless the locale constraint is clearly documented and justified.
The manifest says the skill focuses on industrial/commercial registration information such as legal representative, shareholders, equity penetration, executives, registered/business addresses, phone, and email, and explicitly excludes bidding-behavior traces. In the README, the feature list expands the scope to a '四维扫描框架' including '经营痕迹（D）' and network-density analysis, which goes beyond the manifest's stated limited工商登记层面 scope and suggests broader analytical behavior.
The skill description and instructions are entirely written as a Chinese-only expert persona and do not indicate that users may choose another language. This creates a natural-language policy risk because the skill appears to enforce a specific language/locale without explicit user opt-in or an offered alternative.
The auto-trigger guidance lists only Chinese semantic keywords for activation and does not mention multilingual equivalents or user language choice. This reinforces a Chinese-only activation policy in natural language, which can violate language/locale policy expectations unless clearly justified and opt-in.
The entire example dialog, trigger guidance, and expected output are written as mandatory Chinese-language content, with no indication that users may choose another language or locale. Under the policy rule, a skill that effectively requires a specific language without opt-in is a natural-language policy violation unless the locale restriction is explicitly justified.
The entire skill guidance is written as a Chinese-only operating policy and does not indicate that users may choose another language or locale. Under the policy rule, forcing a specific language without opt-in is a natural-language policy concern unless the locale limitation is explicitly justified.
Line L027 emphatically states the skill '严格不越界' regarding collusion-behavior analysis under Articles 39/40. However, earlier lines describe '经营痕迹（D）', '高度可疑（需核查)', and multi-entity network-density analysis, which are at least partially adjacent to suspicious-behavior screening rather than purely工商登记 matching, creating tension between the declared boundary and the documented behavior.
The primary skill metadata is Chinese, but the clawhub display name, description, and triggers switch to English for one platform variant. This introduces a locale choice implicitly rather than explicitly, and the manifest does not document user opt-in or platform-specific language selection.
No suspicious patterns detected.
