sag
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill metadata references the installation of the
sagbinary via a third-party Homebrew tap (steipete/tap/sag). This is a standard distribution method for command-line tools and is consistent with the skill's stated purpose. - [COMMAND_EXECUTION]: The skill provides instructions for executing the
sagcommand to convert text into speech. Examples show the command being used with specific voice parameters and outputting to local temporary storage (/tmp/voice-reply.mp3). - [CREDENTIALS_UNSAFE]: The skill correctly identifies the need for an
ELEVENLABS_API_KEYand instructs the user to provide it via environment variables, which is a secure practice for managing API credentials. - [DATA_EXFILTRATION]: The tool sends text to the ElevenLabs API to generate audio. This is the intended primary function of the skill and does not involve unauthorized data movement.
- [PROMPT_INJECTION]: The skill's instructions focus on providing a utility for text-to-speech and do not contain any patterns intended to bypass AI safety guidelines or override system behavior.
Audit Metadata