The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates project validation and asset processing by instructing the agent to execute local Node.js scripts, such as scripts/check-spatial-canvas.cjs for spatial integrity and scripts/tts-fish.mjs for narration synthesis. It also leverages the hyperframes-cli through npx for linting, snapshotting, and final video rendering.
[INDIRECT_PROMPT_INJECTION]: The skill architecture includes ingestion points for user-authored planning files, specifically STORYBOARD.md and SCRIPT.md. Furthermore, it documents a subagent dispatch mechanism (DISPATCH) that propagates context to child agents. While these represent surfaces for indirect instructions, the skill defines strict technical contracts and schemas to mitigate accidental behavior overrides.
[EXTERNAL_DOWNLOADS]: The skill references production-standard libraries including GSAP and the Tailwind CSS browser runtime from well-known content delivery networks (e.g., jsdelivr.net). These references are intended for inclusion in the generated HTML compositions rather than direct execution in the agent's host environment.

hyperframes-core