bass-transcribe

Pass

Audited by Gen Agent Trust Hub on Oct 5, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires several Python libraries to perform audio analysis. These include librosa, numpy, pretty_midi, crepe, swift-f0, and basic-pitch. These are standard open-source tools within the audio processing and machine learning community.
  • [COMMAND_EXECUTION]: The skill facilitates the processing of audio files through the execution of Python scripts and CLI commands. It uses a Python interpreter to run transcription pipelines and invokes the basic-pitch command-line tool for polyphonic audio transcription.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided audio files (such as WAV, MP3, and AIFF) which represent an external data ingestion surface. * Ingestion points: Audio data is loaded from user-specified file paths into the processing scripts (e.g., via librosa.load). * Boundary markers: No explicit instruction boundary markers are used when processing audio data, as the input is non-textual. * Capability inventory: The skill can execute Python code, invoke CLI tools, write MIDI files to disk, and interact with the Ableton Live environment through the ableton-mcp skill. * Sanitization: No specific sanitization is performed on the audio input, though the risk of indirect prompt injection via audio signal analysis is negligible.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 5, 2026, 06:46 AM