The Agent Skills Directory

[SAFE]: The skill consists of markdown instructions and configuration files for coordinating frontend interactivity. It does not contain executable code, scripts, or network operations.
[DATA_EXPOSURE]: The skill explicitly advises against security-sensitive practices, such as embedding secrets, signed tokens, or authorization facts in local browser state.
[PROMPT_INJECTION]: The instructions include proactive security advice to prevent injection attacks, such as escaping server-rendered values, using text binding APIs for untrusted content, and avoiding unsafe APIs like x-html for untrusted data.

htmx-interactivity