The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several local utility scripts (new-film.sh, render.mjs, stills.mjs, mux.sh) to automate the motion design workflow. These scripts invoke standard tools such as ffmpeg, npm, and playwright to perform video encoding, dependency management, and frame capture within the project directory.
[EXTERNAL_DOWNLOADS]: Fetches brand guidelines and design tokens from design.the9klabs.com. These resources are managed by the vendor to ensure the animation engine remains synchronized with the 9k design system. The skill also facilitates the installation of standard development dependencies like Playwright and IBM Plex fonts from well-known registries.
[INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by ingesting external references and capturing screenshots of product URLs via browser automation.
Ingestion points: Round 3 of the questionnaire (references/questionnaire.md) and the Playwright-based asset capture in the workflow.
Boundary markers: The instructions in references/engine.md explicitly direct the agent to "take the grammar of the reference, never its content," acting as a logical constraint on the influence of external data.
Capability inventory: The agent has the ability to write project files, execute rendering subprocesses, and automate browser interactions.
Sanitization: External content is primarily processed visually (extracting grammar, shot lengths, and colors) rather than being treated as instructions, reducing the risk of accidental obedience to embedded text.

9k-motion