The Agent Skills Directory

[DATA_EXPOSURE]: Technical guidance for AWS SES integration correctly uses environment variables via craft\helpers\App::env() rather than hardcoding credentials, following secure configuration practices.
[INDIRECT_PROMPT_INJECTION]: The skill documents handling of untrusted data in search results and feeds. It implements defensive patterns such as | striptags, | e('html'), and <![CDATA[...]]> to prevent cross-site scripting (XSS) and other injection attacks.
[EXTERNAL_DOWNLOADS]: Reference files include standard snippets for integrating well-known analytics and consent management platforms (GTM, UserCentrics, CookieBot). These references are neutral, use established domains, and are consistent with the skill's purpose for front-end integration.
[COMMAND_EXECUTION]: Instructions involve standard project-config and asset-building commands using ddev and npm, which are routine operations within a controlled development environment.

craft-site