skills/marswaveai/skills/music/Gen Agent Trust Hub

music

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell commands to interact with the listenhub CLI, manage local configuration via jq, cat, and echo, and utilize utilities like stat for file validation.- [EXTERNAL_DOWNLOADS]: The skill employs curl to retrieve generated music files and stem packages from remote URLs provided by the ListenHub API. These downloads are directed to the current working directory as part of the asset retrieval process.- [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it interpolates user-provided prompts and lyrics into shell commands and processes external media files (audio, images, video).
  • Ingestion points: User-supplied text fields (prompt, lyrics) and local or remote media files (audio, video, image).
  • Boundary markers: The skill uses an explicit confirmation gate via the AskUserQuestion tool before command execution and utilizes double-quoting in shell command construction.
  • Capability inventory: File system writes (echo, mkdir), network downloads (curl), and CLI tool execution (listenhub).
  • Sanitization: Implements file extension checks and size limits (10MB) for local media files.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 06:23 AM