The Agent Skills Directory

[COMMAND_EXECUTION]: The skill directs the agent to run several local scripts for scaffolding components and performing quality audits. These include python3 scripts/scaffold_component.py and various Node.js scripts like verify_states.mjs, axe_audit.mjs, and verify_responsive.mjs located within the project's directory structure.
[INDIRECT_PROMPT_INJECTION]: The skill relies on reading multiple external files from the local repository to define its behavior and output quality.
Ingestion points: .claude/rules/components.md, CLAUDE.md, accessibility/aria-patterns.md, tokens/*.json, and taste/design-taste.md (SKILL.md).
Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within these files.
Capability inventory: The agent has the ability to read files, execute shell commands (e.g., ls), and run local Python/Node.js scripts.
Sanitization: No input validation or sanitization of the content from these project files is mentioned.
[DYNAMIC_EXECUTION]: The skill involves the dynamic generation of a "states harness" HTML file which is then executed and tested using Playwright and automated audit scripts. It also includes a JavaScript snippet for performing visual inspections.
[EXTERNAL_DOWNLOADS]: The documentation suggests the use of npx skills add and npx ux-ui-agent-skills init as corrective actions for a user to take if the design kit is missing, which involves fetching and running packages from a remote registry.

design-component