The Agent Skills Directory

[SAFE]: The skill is a comprehensive toolkit for UI/UX work, including design critique, implementation workflows, and verification steps. It is well-documented and follows a clear, agent-led interaction model with 'investment gates' to ensure user alignment.
[PROMPT_INJECTION]: The skill includes an indirect prompt injection surface as it ingests 'project context' from the user's workspace (e.g., design philosophy, constraints) to guide its UI implementation. This is a primary functional requirement for its task. No boundary markers are explicitly defined to isolate this context, but this is assessed as a standard low-risk factor for design-oriented skills.
[REMOTE_CODE_EXECUTION]: No remote code execution or unauthorized package installations were found. The skill references standard development commands (like build or lint checks) for the agent to use during implementation verification, which is consistent with the primary purpose of frontend development work.
[DATA_EXFILTRATION]: No patterns indicative of data exfiltration or credential harvesting were detected. The skill operates entirely on local project files and its internal reference catalog.
[EXTERNAL_DOWNLOADS]: No external downloads or remote script fetches are present. All design references (Airbnb, Apple, Stripe, etc.) are provided as local markdown files within the skill's package.

tasteful-ui