The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted performance data to inform the generation of new ad creative, which creates a vulnerability surface for indirect prompt injection.
Ingestion points: Performance data provided via CSV, manual text pastes, or API outputs (e.g., Google Ads reports).
Capability inventory: Access to campaign management tools and numerous external generative AI APIs for image, video, and voice generation.
Boundary markers: The instructions do not define clear boundaries or 'ignore' directives for the data being analyzed.
Sanitization: There is no mention of sanitization or structural validation for the ingested metrics and descriptions.
[COMMAND_EXECUTION]: The skill interacts with advertising platforms using local CLI tools.
Evidence: Commands such as node tools/clis/google-ads.js are used to fetch performance data and manage campaign assets.

ad-creative