Security audit
Security checks for vulnerabilities and agentic risk
This is a UI polish checklist with broad activation language, but it does not request sensitive access, execute code, persist changes, or perform behavior outside its stated visual-design purpose.
Install this only if you want an agent to spend time applying a final visual polish pass to UI work. Be aware that its trigger wording is broad, so users who prefer faster prototypes or non-visual implementation work may want narrower activation rules.
The skill declares itself as a core pack item that is 'always active' and the final step before presenting work, which creates an unnecessarily broad trigger scope. In an agent system, over-broad auto-activation can cause the skill to run in contexts where it is irrelevant, override user priorities, consume time, or nudge outputs toward cosmetic changes when the task requires something else.
The instruction to use the skill 'on all visual work' after two other checks is still too broad because it lacks boundaries for task type, project phase, or user request. This can lead to inappropriate invocation across many sessions, making the agent less predictable and potentially interfering with higher-priority instructions or non-polish objectives.
No suspicious patterns detected.
