The Agent Skills Directory

[SAFE]: The skill provides instructions for initializing a design system using legitimate Pencil MCP tools (open_document, set_variables, batch_design, get_screenshot). All operations are scoped to design document manipulation.
[SAFE]: The skill references the official Layui documentation (layui.dev), which is a well-known and established open-source UI library. References to external documentation are used for design token verification.
[INDIRECT_PROMPT_INJECTION]: The skill has a minor attack surface as it ingests user requirements and instructs the agent to consult external documentation to generate design tokens. However, the capabilities are limited to design document variables and layout frames within a controlled environment, and the documentation source is a well-known service.
Ingestion points: User input mentioning "Pencil" and "Layui", and external content from official Layui documentation.
Boundary markers: None explicitly defined.
Capability inventory: mcp__pencil__set_variables and mcp__pencil__batch_design tools for modifying .pen files.
Sanitization: The skill relies on standard agent behavior and does not include specific sanitization instructions for ingested data.

pencil-ui-design-system-layui