The Agent Skills Directory

[SAFE]: The skill is entirely instructional, providing a methodology for developers to improve user experience during asynchronous operations. The provided CSS and HTML snippets in references/patterns.md follow standard industry practices and include appropriate accessibility attributes (e.g., aria-busy, aria-hidden, and prefers-reduced-motion support).
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to inventory asynchronous surfaces by searching (grepping) the local codebase for patterns like fetch(, axios, and useQuery. While this involves ingesting untrusted data (the project's code), the capabilities are limited to generating a markdown audit table and standard UI components. The risk is considered negligible as there are no network exfiltration or remote code execution vectors present.
[SAFE_CONTEXT]: All logic and code examples are self-contained. The skill does not attempt to download external scripts, install unverified packages, or access sensitive system files like credentials or SSH keys.

loading-states