The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the execution of the sag binary to perform its primary function of text-to-speech generation.
[EXTERNAL_DOWNLOADS]: The skill metadata references the installation of the sag utility via a Homebrew tap. Homebrew is a well-known and established service for package distribution.
[PROMPT_INJECTION]: The usage instructions in SKILL.md for the 'Chat voice responses' feature demonstrate a vulnerability to indirect prompt injection. The example command sag -v Clawd -o /tmp/voice-reply.mp3 "Your message here" interpolates user-provided text directly into a shell command without escaping or sanitization.
Ingestion points: The user-provided message content described in the bash example in SKILL.md.
Boundary markers: No delimiters or safety instructions are provided to the agent to prevent the interpretation of malicious content within the user message.
Capability inventory: The agent uses shell command execution via the sag tool.
Sanitization: There are no instructions for sanitizing, validating, or escaping user input before it is used in the command line.

sag