The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references documentation, setup guides, and image assets from the vendor's official infrastructure, including the inference.sh domain and the inference-sh/skills GitHub repository. These are recognized as legitimate vendor resources.
[COMMAND_EXECUTION]: The instructions involve the use of the belt CLI tool for authentication and application execution. The commands are standard for the tool's intended purpose and are scoped to the vendor's ecosystem.
[INDIRECT_PROMPT_INJECTION]: The skill serves as a processing point for user-supplied text to be converted into speech, creating a potential surface for indirect injection.
Ingestion points: Text input passed to the belt app run command via the --input JSON parameter in SKILL.md examples.
Boundary markers: The input is encapsulated within structured JSON objects, providing logical separation between data and command parameters.
Capability inventory: The skill uses the Bash tool to execute the belt CLI for speech generation.
Sanitization: Input sanitization or filtering of the text content is not explicitly described in the skill documentation.

elevenlabs-tts