The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data that is passed directly to the belt CLI, creating a surface for indirect instructions to influence the model's output or behavior.
Ingestion points: User-provided strings for voice_script, video_prompt, and voice_prompt parameters in SKILL.md examples.
Boundary markers: None identified. There are no instructions to the agent to treat these inputs as untrusted data or to ignore embedded commands.
Capability inventory: The skill uses Bash(belt *) to execute commands on the system.
Sanitization: No sanitization or validation of the input strings is performed before they are interpolated into shell commands.
[COMMAND_EXECUTION]: The skill requires the Bash tool to execute belt CLI commands, which is the primary mechanism for its functionality.
[EXTERNAL_DOWNLOADS]: The skill references external resources and installation scripts.
Fetches installation instructions from the inference-sh GitHub repository.
Instructs the user to add the belt-sh/cli and other inference-sh skills via npx.

p-video-avatar