The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted user data, including video briefs, images, videos, and product URLs, to generate instructions for video production models. Ingestion points: User-provided text and media are analyzed and incorporated into storyboards and prompts in SKILL.md and storyboard-route.md. Boundary markers: The skill does not explicitly instruct the agent to use delimiters or ignore-embedded-instruction warnings when processing external data. Capability inventory: The skill can invoke video_generate (resource-intensive generation), video_stitch (file manipulation), and various media analysis tools. Sanitization: No explicit content sanitization or validation is required before processing user inputs, though safety filter handling is mentioned as a fallback.
[SAFE]: The skill exhibits no malicious patterns such as obfuscation, hardcoded credentials, or unauthorized remote code execution. It relies on standard platform tools and follows logical workflows for its stated purpose.

generating-videos