The Agent Skills Directory

[SAFE]: The skill demonstrates safe behavior throughout its files. No signs of obfuscation, hardcoded credentials, unauthorized network operations, or malicious intent were found during the analysis.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided natural language descriptions to generate design systems, creating a standard interaction surface.
Ingestion points: Descriptions are ingested as command-line arguments by scripts/match-trend.ts to identify design trends.
Boundary markers: None are explicitly used to separate user input from internal instructions.
Capability inventory: The skill generates design tokens, CSS variables, and Tailwind configurations based on the analysis.
Sanitization: Input strings are processed for keyword matching and score calculation within the script's logic, without being passed to a shell or dynamic evaluator.
[COMMAND_EXECUTION]: The skill uses local TypeScript scripts (scripts/*.ts) to handle trend matching and code generation. These scripts are intended to be executed via ts-node as part of the primary workflow and do not exhibit dangerous command injection patterns or elevated privilege requirements.
[DATA_EXPOSURE]: The generator scripts reference a design catalog file at a hardcoded relative path (../../../../website/design-catalog/gallery-sources.json). This appears to be a functional dependency on the vendor's internal project structure rather than an attempt to access sensitive system files.

design-system-generator