The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided text which is then interpolated into JSON structures passed to the belt CLI via Bash. While this is standard for TTS tools, it creates a surface where malformed input could potentially influence command execution if not properly escaped by the agent platform. Ingestion points: Processed text fields such as text and voice_script in SKILL.md examples. Boundary markers: The examples do not demonstrate the use of specific delimiters to isolate user-provided content from the command structure. Capability inventory: The skill utilizes Bash(belt *) which allows execution of CLI commands on the host system. Sanitization: There are no explicit instructions or middleware steps described for sanitizing the input text before interpolation.
[EXTERNAL_DOWNLOADS]: The skill directs the agent to fetch installation documentation and potential configuration data from external GitHub repositories. Fetches setup guidelines from the inference.sh official GitHub repository at https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.

text-to-speech