skills/b-open-io/prompts/voice-clone/Gen Agent Trust Hub

voice-clone

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/voice-clone.ts script utilizes Bun.spawn to execute external command-line tools including ffmpeg (for audio trimming, normalization, and noise removal) and yt-dlp (for downloading audio from video platforms). These tools are executed using argument arrays, which prevents shell command injection.
  • [EXTERNAL_DOWNLOADS]: The skill downloads audio files from user-specified URLs and communicates with the well-known ElevenLabs service at api.elevenlabs.io to upload voice samples and generate speech. These operations are core to the skill's functionality.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external URLs and local audio files. While this establishes a theoretical attack surface for indirect prompt injection (e.g., via malicious file metadata), the skill does not currently parse text-based instructions from these sources to drive agent behavior, and the ingestion is required for the intended purpose of voice cloning.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 04:00 PM