The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill documentation explicitly recommends that users on macOS and Linux install the required CLI tool by running curl -fsSL https://edgespeak.com/install.sh | sh. This pattern is highly dangerous as it executes arbitrary code from a remote server without verification or integrity checks.
[COMMAND_EXECUTION]: The skill provides examples for using the edgespeak-cli speech command where user-controlled text is placed inside double quotes (e.g., edgespeak-cli speech "<text>"). This structure is vulnerable to shell injection if the input text contains characters like backticks, semicolons, or dollar signs, allowing for arbitrary command execution on the host system.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted user-provided text for speech synthesis.
Ingestion points: The text to be synthesized is provided by the user and passed to the speech command in SKILL.md.
Boundary markers: No boundary markers or special delimiters are used to isolate the user text from the command structure, other than standard shell quotes.
Capability inventory: The skill uses shell execution via edgespeak-cli and provides file system write access via the -o output flag.
Sanitization: There is no evidence of input sanitization or filtering to prevent the injection of shell control characters or command sequences.

edgespeak-broadcast