The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill utilizes the npx ui-skills get <skill-slug> command to download and execute code from the npm registry at runtime. This pattern allows for the execution of external, unversioned scripts from third-party sources without integrity verification.
[EXTERNAL_DOWNLOADS]: The process of fetching design 'lenses' involves downloading content from various external npm accounts (e.g., jakubkrehel, ibelick, pbakaus, wshobson). These downloads occur dynamically during skill execution based on the specific design needs identified by the agent.
[COMMAND_EXECUTION]: The skill instructs the agent to use the system open command to launch generated HTML files in a browser, which involves direct interaction with the host operating system's command shell.
[INDIRECT_PROMPT_INJECTION]: The skill processes external design documentation and codebase metadata to 'ground' the mockup generation, which exposes the agent to potentially malicious instructions embedded in project files.
Ingestion points: Reads meta/DESIGN.md and scans existing project files for styling facts and CSS tokens.
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the ingested design files.
Capability inventory: Includes file system write access, execution of system commands (open), and remote code execution (npx).
Sanitization: The skill does not implement any validation or sanitization routines for the design language facts retrieved from the project environment.

generate-mockups