The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection. Ingestion points include untrusted data such as product catalogs, blog posts, and site metadata entering the agent context via site-profile and catalog tools (references/ecommerce/recommend-ecommerce-strategy.md). Boundary markers or instructions to ignore embedded commands are absent in the recipes that interpolate this data. Capability inventory includes significant write operations across Wix business solutions, such as binding domains, configuring payment onboarding, and scheduling social media posts (references/domains/domain-search-purchase-and-connect.md, references/get-paid/how-to-setup-wix-payments.md, references/marketing/generate-and-publish-marketing-plan.md). Explicit sanitization or filtering of external content is not defined in the instructions.
[EXTERNAL_DOWNLOADS]: The documentation references official Wix NPM packages (@wix/sdk) and well-known media/AI services (unsplash.com, runware.ai) for content generation and media management.
[COMMAND_EXECUTION]: The skill provides curl templates for the agent to execute against Wix's administrative REST APIs at wixapis.com and manage.wix.com.

wix-manage