text-to-speech

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface
  • Ingestion points: User-provided text is interpolated into the --input argument for the belt app run command in SKILL.md.
  • Boundary markers: The skill uses JSON formatting (e.g., {"text": "..."}) to encapsulate user input, which provides some structural separation but no explicit instructions to the model to ignore embedded commands.
  • Capability inventory: The skill utilizes the belt CLI tool (Bash) to execute audio generation apps.
  • Sanitization: There is no evidence of input validation or sanitization for the text strings before they are sent to the TTS models.
  • [EXTERNAL_DOWNLOADS]: The skill points to an external CLI installation script at https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md and recommends adding related skills via npx. These resources are hosted by the vendor organization and are necessary for the skill's advertised functionality.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 01:25 PM
Security Audit — agent-trust-hub — text-to-speech