gpt-image

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of external components from the belt-sh and inference-sh repositories using the npx package runner.
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute belt CLI commands for user authentication and model inference tasks.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines an interface that ingests external data in the form of image and mask URLs for processing by the image model.
  • Ingestion points: Found in the 'images' and 'mask' parameters defined in SKILL.md.
  • Capability inventory: The skill has the ability to execute shell commands via the Bash tool to run the 'belt' application.
  • Boundary markers: The skill uses structured JSON inputs for parameters, which helps distinguish between prompt instructions and metadata.
  • Sanitization: No explicit sanitization of the remote image content is performed within the skill definition itself.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — gpt-image