music-prompter

Pass

Audited by Gen Agent Trust Hub on Oct 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is well-structured and focuses on its primary purpose of music generation. It provides detailed instructions for constructing musical prompts and managing API interactions without attempting to bypass safety filters or perform unauthorized operations.
  • [EXTERNAL_DOWNLOADS]: The skill references the official GitHub repository for the SandBase CLI (github.com/sandbaseai/cli) for setup instructions. This is a vendor-owned resource consistent with the skill's authorship and purpose.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where it ingests untrusted data from user musical briefs and external API responses from the sandbase_run_get tool.
  • Ingestion points: User-supplied descriptions in SKILL.md and runtime data from the SandBase API.
  • Boundary markers: None explicitly defined in the instructions.
  • Capability inventory: Access to the sandbase_* toolset for API interaction and local audio processing tools for clip assembly.
  • Sanitization: No specific sanitization steps are documented for processing the musical briefs before they are incorporated into prompts.
  • [COMMAND_EXECUTION]: The skill mentions joining audio clips locally using "deterministic audio tools." This involves standard execution of local utilities (such as ffmpeg) to fulfill the user's request for multi-part music pieces.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 4, 2026, 11:45 AM