ai-music-generation

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references external resources for tool installation and documentation, primarily targeting GitHub and the inference.sh domain. These are standard references for the skill's functionality.
  • [COMMAND_EXECUTION]: The skill utilizes a specific CLI tool, belt, to execute music generation tasks. The execution environment is restricted via the allowed-tools: Bash(belt *) configuration in the frontmatter, limiting the scope of executable commands to the intended tool.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes natural language prompts provided by the user to generate music. This represents a standard surface for indirect prompt injection common to generative AI interfaces.
  • Ingestion points: Prompts and lyrics are ingested via the --input JSON argument in belt app run commands (SKILL.md).
  • Boundary markers: No explicit delimiters are used in the instruction examples for user prompts.
  • Capability inventory: The skill is capable of executing the belt CLI tool to interact with remote inference APIs (SKILL.md).
  • Sanitization: No specific sanitization or filtering of prompt content is described in the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — ai-music-generation