mm-music-expert
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The instructions in SKILL.md direct the agent to execute a shell command that incorporates user-supplied lyrics and prompts (e.g.,
python scripts/generate_music.py --lyrics "<final_lyrics>"). This introduces a potential command injection surface if special shell characters are not properly handled by the execution environment. Additionally, the agent is tasked with determining the output filename from context, which could lead to path traversal if not restricted. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input which is then used as parameters for a generation script.
- Ingestion points: User-provided music lyrics and style prompts collected during the interactive workflow in SKILL.md.
- Boundary markers: The workflow uses markdown structures but lacks explicit delimiters or instructions to ignore potential injection attempts within the lyrics or prompts.
- Capability inventory: The skill executes a local Python script (scripts/generate_music.py) which performs network operations (API calls to api.minimaxi.com) and file system writes (saving audio files to the ./audio/ directory).
- Sanitization: There is no evidence of sanitization or escaping of user-supplied text before it is interpolated into shell commands or sent to the external API.
Audit Metadata