dialogue-audio

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions demonstrate interpolating user-provided text prompts directly into shell commands using the belt CLI (e.g., belt app run falai/dia-tts --input '{"prompt": "..."}').
  • Ingestion points: The prompt field within the JSON input for the belt CLI tool in SKILL.md.
  • Boundary markers: Absent. The examples use single quotes to wrap the JSON payload but do not suggest escaping for the content within the prompt string.
  • Capability inventory: The skill is granted Bash(belt *) access, allowing it to execute the belt command-line utility.
  • Sanitization: Absent. There is no mention of sanitizing or escaping the user's input before it is passed to the shell command.
  • [COMMAND_EXECUTION]: The skill's primary functionality relies on executing the belt CLI tool to run remote applications like falai/dia-tts, infsh/video-audio-merger, and infsh/media-merger. While these are the intended tools for the skill's purpose, they represent a significant capability for executing operations in a shell environment.
  • [EXTERNAL_DOWNLOADS]: The documentation encourages the installation of external tools and resources, such as the belt-sh/cli package via npx and fetching installation instructions from a remote GitHub repository (inference-sh/skills).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — dialogue-audio