The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute framework-specific CLI tools to ensure quality and compliance.
Evidence: Use of npx hyperframes lint, npx hyperframes validate, and npx hyperframes inspect in the output checklist.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied creative briefs and design requirements to generate visual content. This represents a standard attack surface where malicious instructions in a user prompt could attempt to influence the generated code or agent behavior.
Ingestion points: User requests for video compositions (e.g., brand intros, title cards).
Boundary markers: The skill employs a HARD-GATE and strict internal checklists to ensure the agent follows the design spec over the user prompt, which acts as a mitigating control.
Capability inventory: Generates HTML/CSS/JS (GSAP) and executes local CLI tools (npx hyperframes).
Sanitization: Not explicitly defined in this orchestration file, as it relies on the downstream hyperframes core and animation modules for final code output.
[DYNAMIC_EXECUTION]: The skill involves the generation of executable animation scripts (GSAP) and HTML structures based on high-level templates and design system rules.
Evidence: Instructions for writing static CSS and then adding GSAP tweens (gsap.from, gsap.fromTo) to animate elements.

general-video