The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to dynamically adapt its baseline configuration values (DESIGN_VARIANCE, MOTION_INTENSITY, VISUAL_DENSITY) based on explicit user chat prompts, creating a surface where untrusted input influences the agent's logic.
Ingestion points: SKILL.md (Section 1) explicitly directs the agent to "ALWAYS listen to the user" to override hardcoded baseline values.
Boundary markers: The skill lacks specific delimiters or "ignore embedded instructions" warnings to separate user-provided configuration overrides from the agent's core instructions.
Capability inventory: The skill enables the agent to generate and potentially execute React/Next.js code, as well as install third-party Node.js packages (npm install).
Sanitization: There are no defined validation or sanitization protocols for the dynamic configuration values provided by the user.
[EXTERNAL_DOWNLOADS]: The skill references external resources for upstream updates and placeholder assets.
Source: Upstream repository at https://github.com/Leonxlnx/taste-skill is referenced in the metadata.
Source: Suggests using the well-known placeholder service picsum.photos for image assets in generated code.

design-taste-frontend-v1