The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing external data from user-supplied product URLs and listing text (Gate 1). This content is interpolated into prompts for generating scripts and visual assets. While this is a standard functional requirement, it creates a surface for indirect prompt injection where instructions hidden on a product page could influence the agent's behavior.
Ingestion points: The 'Gate 1 — product read' section in SKILL.md specifies sourcing data from external URLs, listing text, or images.
Boundary markers: Absent. The instructions do not define specific delimiters or warnings to ignore instructions embedded in the fetched product data.
Capability inventory: The skill uses arcads_generate_image, arcads_generate_video_seedance_20, arcads_get_upload_url, and arcads_register_image for asset generation and management.
Sanitization: Absent. The skill summarizes and verifies facts but does not perform explicit escaping or filtering of the external content before prompt construction.

pixar-ad