Security audit
Security checks for vulnerabilities and agentic risk
This is a structured UI/design critique skill with broad trigger wording but no executable code, persistence, credential handling, or hidden data access.
Install this if you want a strict, opinionated design-review workflow with scoring and benchmark comparisons. Be aware that broad trigger phrases may make it activate for general UI feedback, and only provide screenshots or URLs you are comfortable having reviewed.
The skill metadata includes broad aliases such as “设计评审”, “UI 审查”, “UX 评审”, and “design review”, which are common phrases users may use in ordinary conversation. This can cause the skill to activate unexpectedly and steer general design-advice requests into this skill’s rigid workflow, creating prompt-routing and scope-confusion risk rather than direct code-execution risk.
The activation scenario says the skill should engage whenever a user asks for professional evaluation or improvement suggestions on an existing interface, but that overlaps heavily with normal assistant behavior for general design feedback. Because the boundary is not clearly constrained, the orchestrator may invoke this skill when the user did not intend a formal scoring framework, leading to unintended tool/skill selection and overreach.
The dedicated trigger list contains generic phrases like ‘优化建议’, ‘改进建议’, ‘这个页面有什么问题’, and ‘review UI’, which are high-frequency phrases across many unrelated product and design conversations. In a skill-routing system, such high-collision triggers increase accidental invocation and can crowd out more appropriate skills or baseline assistant handling.
No suspicious patterns detected.
