The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional content, accessibility checklists, and static code examples for UI development. No executable scripts, network requests, or system-level operations are performed by the skill itself.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a reference library for generating user interface code based on natural language requests.
Ingestion points: The agent ingests user requirements for form controls and applies the patterns found in SKILL.md and references/pattern.md.
Boundary markers: The documentation uses clear headers and structural sections to define implementation steps, though no explicit runtime delimiters are defined for user-provided labels or values.
Capability inventory: The skill is limited to providing HTML snippets, Mermaid diagrams, and design tokens; it does not request or utilize tools for file system modification, network communication, or shell execution.
Sanitization: The instructions prioritize native HTML5 controls and semantic ARIA attributes, which minimizes the attack surface compared to complex custom implementations.

slider