Security audit
Security checks for vulnerabilities and agentic risk
This is a coherent Chinese-language judging criteria skill with a local checker and no evidence of hidden access, persistence, network use, or destructive behavior.
Before installing, note that the skill is mainly Chinese-language and may activate on broad AUT/UIBC-related terms. It appears suitable for its stated judging workflow, but English-only users may need translation and maintainers should consider narrowing trigger wording.
The entire agent-facing instruction file is written in Chinese, including the title and operational guidance, with no indication that language selection is optional or that the skill is intended only for Chinese-speaking users. Under the stated policy, forcing a specific language without user opt-in is a natural-language policy violation.
The README describes SKILL.md and 判据_AUT.md as the AI entry point and authoritative answer source, and both descriptions are only in Chinese. Although line L43 adds a brief English summary for the README, the actual skill-entry materials are presented as Chinese-only, which can amount to a language-policy constraint without explicit user opt-in or documented justification.
The metadata description and read_when triggers include broad generic terms such as 'UIBC', '自动驾驶', '迁移判定', and '责任归属' without requiring a narrowly scoped AUT-domain judging request. This can cause the skill to activate during ordinary discussion and inject domain-specific judging rules where they were not requested, increasing the chance of irrelevant guidance, unintended authority escalation, or workflow hijacking.
The '何时用' section repeats broad and ambiguous trigger guidance, allowing activation on common topic mentions rather than a clearly bounded request for this skill's specialized function. In an agent system, this increases over-triggering risk, which can misroute tasks, override more appropriate skills, and expose users to authoritative-sounding but contextually inappropriate outputs.
The module docstring and all user-facing CLI messages are written only in Chinese, and the script is described as producing executable judgment output directly. For an all-file-types language policy check, this can indicate a fixed language/locale behavior without any opt-in or alternative language path.
