speak-response

Pass

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: SAFE
Full Analysis
  • COMMAND_EXECUTION (SAFE): The skill performs audio synthesis and playback by calling local shell and Python scripts. It uses proper variable quoting in bash to prevent command injection from user-provided text.
  • EXTERNAL_DOWNLOADS (SAFE): The skill requires cloning the Qwen3-TTS repository and downloading weights from Hugging Face. These are from highly reputable sources in the AI field and are integral to the skill's primary function.
  • DATA_EXFILTRATION (SAFE): Documentation provides an example for using the OpenAI Whisper API for transcription, which is a legitimate and transparent use case. No clandestine data exfiltration was observed.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 21, 2026, 05:49 PM