music-timbre
Warn
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill interpolates the user-provided
<audio_file_path>directly into a shell command:python3 -m music_analyzer timbre "<audio_file_path>". This creates a risk of command injection if the input contains shell metacharacters (e.g.,;,&,|,`,$()) that could terminate the intended command and execute arbitrary code. - [INDIRECT_PROMPT_INJECTION]: The skill processes external audio files which may contain untrusted metadata (e.g., ID3 tags, artist/title fields) that could influence the agent's behavior if the analysis results are reflected into the prompt.
- Ingestion points: Audio file content and metadata via the
<audio_file_path>input inSKILL.md. - Boundary markers: None present to delimit or sanitize output from the
music_analyzertool. - Capability inventory: Shell command execution using
python3inSKILL.md. - Sanitization: The skill includes a text instruction to "Validate the audio file path", but lacks technical enforcement or escaping of the path before shell interpolation.
Audit Metadata