skills/magentosh/skills/gpt-image/Gen Agent Trust Hub

gpt-image

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references external documentation and installation scripts hosted on the inference.sh official website and its public GitHub repository for setting up the required CLI environment.
  • [COMMAND_EXECUTION]: Functional capabilities are provided via the belt command-line interface, which is used for authentication (belt login) and executing remote image processing applications (belt app run).
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied text for prompts, image URLs, and masks which are interpolated into shell commands. While no explicit sanitization is shown in the examples, the tool's scope is restricted to image generation parameters and is managed within the belt CLI ecosystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 06:03 PM
Security Audit — agent-trust-hub — gpt-image