The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use npx hyperframes for previewing, linting, and rendering video compositions. These commands are localized to the project directory and are standard for the described video production workflow.
[EXTERNAL_DOWNLOADS]: Utilizes npx to execute the hyperframes package, which may be fetched from the public npm registry if not present in the local environment. This is a common practice for Node.js-based tooling.
[INDIRECT_PROMPT_INJECTION]: The skill processes HTML, CSS, and GSAP JavaScript files within the compositions/ directory.
Ingestion points: Reads scene source files (HTML/CSS/JS) from the local file system (e.g., compositions/scene-name.html).
Boundary markers: None explicitly defined in the skill instructions to separate file content from agent instructions.
Capability inventory: Execution of shell commands via npx for rendering and linting.
Sanitization: No explicit sanitization or validation of the input HTML/JS is mentioned; however, the skill focuses on structured scene building rather than arbitrary data processing.

hyperframes-scene-builder