gpt-image
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill utilizes a pinned version of the
mcp-openai-propackage (v2026.10.19.3) to provide its core service functionality. It also references official documentation on the vendor's GitHub repository for advanced usage. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied prompts for image generation and editing, which introduces a standard indirect injection surface.
- Ingestion points: User prompts and reference data provided during the interaction (SKILL.md).
- Boundary markers: No specific delimiters or safety warnings for tool input parameters are defined in the instructions.
- Capability inventory: The skill utilizes tools for image generation, editing, and task status polling (
gpt_image_generate,gpt_image_edit,gpt_image_get_task). - Sanitization: No explicit sanitization or filtering of user input is specified before it is sent to the image processing tools.
Audit Metadata