ai-voice-cloning

Fail

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION] (HIGH): The skill documentation includes instructions to install a CLI tool using a piped shell execution pattern.
  • Evidence: curl -fsSL https://cli.inference.sh | sh in the Quick Start section of SKILL.md.
  • Risk: This method executes code directly from a remote server without verification, allowing the owner of the domain to execute arbitrary commands on the user's machine.
  • [EXTERNAL_DOWNLOADS] (MEDIUM): The skill relies on software from a non-trusted external source.
  • Evidence: All tools and installation scripts are hosted on inference.sh.
  • Context: The domain inference.sh is not on the established list of trusted organizations or repositories.
  • [PROMPT_INJECTION] (LOW): The skill exhibits an indirect prompt injection surface by processing untrusted text data through a powerful CLI tool.
  • Ingestion points: Untrusted text is passed to the text field in infsh app run commands (e.g., in SKILL.md).
  • Boundary markers: Absent; there are no delimiters or instructions to prevent the agent from interpreting instructions embedded within the text to be voiced.
  • Capability inventory: The skill has access to Bash(infsh *), which allows for broad command execution.
  • Sanitization: Absent; no sanitization or escaping of the input text is mentioned.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 18, 2026, 11:05 AM