The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use the infsh CLI tool for managing and running Google Veo video generation applications. This includes commands for authentication and application execution.
[EXTERNAL_DOWNLOADS]: The skill provides a link to installation instructions for the required CLI tool, hosted on the author's official GitHub repository (inference-sh/skills).
[PROMPT_INJECTION]: The skill processes user-supplied text prompts which are passed as input to the infsh command. This creates a surface for indirect prompt injection where malicious instructions embedded in a prompt could attempt to influence the behavior of the video generation model.
[INDIRECT_PROMPT_INJECTION_EVIDENCE]:
Ingestion points: User input provided to the --input flag in infsh app run commands.
Boundary markers: Uses a JSON structure for input arguments, providing structural separation but not preventing semantic injection.
Capability inventory: The infsh tool executes network requests to the inference.sh API and manages local application state.
Sanitization: No explicit sanitization or validation of the prompt content is defined within the skill's instructions.

google-veo