The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as customer reviews and competitor ad library content. It mitigates risk with explicit instructions to treat this content as data only and ignore any embedded directives. Ingestion points are identified in CREATIVE-ROADMAP.md.
[COMMAND_EXECUTION]: The documentation outlines workflows for advanced creative production using CLI tools like ffmpeg and Playwright. These are functional requirements for generating video ads described in NATIVE-REVEAL-VIDEO.md.
[SAFE]: The provided HTML review template (assets/creative-review-template.html) includes security features to prevent data exfiltration, specifically by blocking the loading of remote or protocol-relative image assets.

ad-creative