The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the edgespeak-cli tool to perform text segmentation, check application status, and manage licensing. It incorporates safety instructions that explicitly forbid the AI agent from executing the installation script itself, requiring the user to perform the installation.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted text data (transcripts, captions, or dictation) which could contain malicious instructions designed to influence the behavior of the agent or the underlying model.
Ingestion points: The skill accepts raw text via the --text argument or file content via the --file and --transcript arguments in SKILL.md.
Boundary markers: The skill does not implement delimiters or specific instructions to disregard potentially malicious content within the processed text.
Capability inventory: The tool is limited to executing edgespeak-cli for text processing tasks and does not have built-in network exfiltration capabilities beyond standard CLI licensing checks.
Sanitization: There is no documentation of input sanitization or validation performed on the text before it is passed to the CLI tool.

edgespeak-segment