The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the AI agent to dynamically adapt global configuration variables (DESIGN_VARIANCE, MOTION_INTENSITY, VISUAL_DENSITY) based on explicit user requests in chat prompts, which then drive the logic for code generation.
Ingestion points: User-provided instructions in chat prompts are directly used to override the ACTIVE BASELINE CONFIGURATION in SKILL.md.
Boundary markers: The skill lacks explicit delimiters or warnings to treat user-provided configuration overrides as untrusted data.
Capability inventory: The agent is authorized to generate functional React/Next.js components and output shell commands for package installation (npm install).
Sanitization: No instructions are provided for validating or sanitizing user-provided input before it influences the code generation logic.
[SAFE]: The skill references several well-known and trusted external libraries and services.
Fetches placeholder images from the well-known picsum.photos service.
Instructs the agent to use established packages such as framer-motion, lucide-react, zustand, tailwindcss, gsap, and three from official registries.

design-taste-frontend-v1