telegram-voice-to-voice-macos

Pass

Audited by Gen Agent Trust Hub on Feb 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple system commands including say, ffmpeg, defaults, and yap via shell scripts (transcribe_telegram_ogg.sh, tts_telegram_voice.sh).
  • [EXTERNAL_DOWNLOADS]: The documentation references an external tool yap (https://github.com/finnvoor/yap) required for functionality. While the skill doesn't download it automatically, it relies on this third-party CLI.
  • [DATA_EXPOSURE]: The skill reads from and writes to specific paths in the user's home directory (~/.openclaw/media/inbound/ and ~/.openclaw/workspace/voice_out), which is necessary for its stated purpose of audio processing.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes external audio data.
  • Ingestion points: Reads .ogg files from ~/.openclaw/media/inbound/ in scripts/transcribe_telegram_ogg.sh.
  • Boundary markers: None identified in the provided scripts or instructions.
  • Capability inventory: Executes subprocesses (yap, ffmpeg, say, defaults) and performs file writes/deletes via shell scripts.
  • Sanitization: No specific sanitization or validation of the inbound audio content or the transcribed text before it is used for reply generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 26, 2026, 03:08 PM