The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the kokoro CLI tool via shell commands. While legitimate for the skill's purpose, constructing shell commands like kokoro speak "Text to read" using external or user-provided text creates a potential command injection surface if the text contains shell metacharacters.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the kokoro-cli package and subsequent setup (likely downloading model weights) using the uv package manager. This is a standard procedure for this tool but involves fetching code and data from external repositories.
[INDIRECT_PROMPT_INJECTION]: The skill processes text from the agent's context to generate audio. If the source of this text is untrusted (e.g., content from a website), it could contain instructions intended to manipulate the agent, though the impact is limited to the audio output and CLI parameters.
Ingestion points: Text passed to the kokoro speak command (SKILL.md).
Boundary markers: Not present.
Capability inventory: Subprocess execution of the kokoro CLI tool (SKILL.md).
Sanitization: None specified; relies on the agent's internal handling of shell arguments.

read-aloud