The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to ingest external data such as titles, metrics, and JSON props to dynamically generate video frames. This creates a potential surface for indirect prompt injection where malicious instructions could be embedded in the data processed by the agent.
Ingestion points: The MainComposition component props and the --props CLI flag described in SKILL.md.
Boundary markers: No delimiters or explicit instructions to ignore embedded commands are present in the provided React templates.
Capability inventory: The skill utilizes shell command execution via npx and filesystem writes for the final video output.
Sanitization: No input validation or sanitization logic is implemented for the data-driven content in the example code.
[COMMAND_EXECUTION]: The skill provides instructions for executing shell commands to preview and render video content.
Evidence: Shell commands npx remotion studio and npx remotion render MainVideo output.mp4 are used for video processing.
[EXTERNAL_DOWNLOADS]: The skill utilizes the remotion package, an established framework for programmatic video, which is fetched and executed via the npx command at runtime.

programmatic-video