Security audit
Security checks for vulnerabilities and agentic risk
This looks like a UI component generator, but it asks for broad write and command execution authority without tight scope or clear user controls.
Review this before installing if you do not want a UI helper skill to have general command execution or file-writing authority. It is safest to use it only for generating code in chat or saving explicitly named HTML files, and to avoid letting it run shell commands except for a deliberate local preview you requested.
The skill declares exec despite its stated purpose being generation of single-file HTML/CSS components, which does not inherently require shell execution. Unnecessary command execution materially increases attack surface because an agent could create, open, or run local commands on the user system under the pretext of normal skill usage.
The skill claims to be a narrowly scoped local HTML/CSS component generator, but the documentation broadens behavior into generic create/query/modify/delete, import/export, and runtime configuration flows. That mismatch weakens least-privilege expectations and can enable an agent to justify actions outside the advertised UI-generation scope, increasing the chance of unsafe file or command operations.
The skill includes write and exec capabilities and even suggests local commands to open files, but it does not prominently warn that files may be created or commands may run on the user machine. Missing disclosure undermines informed consent and increases the risk of unintended local side effects from ordinary-seeming component-generation requests.
The trigger condition covers broad development activities such as code generation, programming assistance, debugging, testing, and deployment, far beyond UI component generation. Overbroad invocation criteria make accidental activation more likely and can route unrelated requests into a skill that has write/exec capabilities, increasing misuse risk.
The file states the skill is fully local and does not involve external APIs, but elsewhere requires an LLM API and notes network/model dependence. This inconsistency can mislead users about data flow and trust boundaries, causing them to provide sensitive input under false assumptions about locality and external exposure.
The display name, summary, description, examples, and operational guidance are all written in Chinese, but the document does not state that the skill is China-specific or provide users an option to choose another language. This can violate language/locale policy when a skill effectively forces a specific language without user opt-in.
