The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to interpret user chat prompts to adapt design variables and generate corresponding UI code, which creates a potential surface for indirect prompt injection.
Ingestion points: The agent is instructed to dynamically adapt its configuration based on user chat prompts (Section 1: ACTIVE BASELINE CONFIGURATION).
Boundary markers: There are no explicit delimiters or instructions to treat user input as untrusted data when generating code or commands.
Capability inventory: The skill can generate npm install commands and write complex React/Next.js code involving multiple external libraries.
Sanitization: No input validation or sanitization logic is provided to ensure user-requested libraries or components are safe.
[COMMAND_EXECUTION]: The instructions mandate the generation of npm install commands whenever a 3rd-party library (like framer-motion or zustand) is required but missing from package.json (Section 2: DEFAULT ARCHITECTURE & CONVENTIONS).
[EXTERNAL_DOWNLOADS]: The skill promotes the use of various external UI libraries and resources.
Recommends standard frontend packages including framer-motion, lucide-react, zustand, @phosphor-icons/react, and @radix-ui/react-icons.
Utilizes picsum.photos as a reliable source for placeholder image assets (Section 7: AI TELLS).
[DYNAMIC_EXECUTION]: The core purpose of the skill is the runtime generation of executable frontend code (React, Next.js, Framer Motion, GSAP, ThreeJS) that is intended to be run in the user's development environment.

design-taste-frontend