The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references documentation and installation guidelines hosted on the inference-sh GitHub repository. These references are used to set up the necessary environment for the skill's primary function.
[COMMAND_EXECUTION]: The skill makes extensive use of the 'belt' CLI tool to run audio generation models (e.g., falai/dia-tts). These commands are standard operations for interacting with the inference.sh infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill takes user-defined text prompts and passes them to a text-to-speech model. While this involves interpolating untrusted text into a command-line payload, the target is a specialized audio generation model on a managed platform, which effectively isolates the agent from potential prompt injection attacks affecting the agent itself.

dialogue-audio