gpt-image

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references and facilitates the download of configuration and installation content from a remote repository at raw.githubusercontent.com/inference-sh/skills.
  • [COMMAND_EXECUTION]: The skill requires the installation and execution of external CLI tools (belt-sh/cli) using npx and provides instructions for running applications via the belt command-line interface.
  • [INDIRECT_PROMPT_INJECTION]: The skill allows the agent to ingest external URLs for images and masks, which could potentially contain instructions intended to influence the model's behavior.
  • Ingestion points: The images array and mask parameters in SKILL.md accept external URLs.
  • Boundary markers: No delimiters or explicit instructions to ignore embedded content are present.
  • Capability inventory: The skill utilizes the Bash(belt *) tool to execute shell commands.
  • Sanitization: No sanitization or validation of the external content is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 11:45 AM
Security Audit — agent-trust-hub — gpt-image