gpt-image-edit
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documents the execution of the
runcomfyCLI tool to perform image editing. It provides structured examples of passing JSON inputs, including prompts and image URLs, to the command-line interface. - [INDIRECT_PROMPT_INJECTION]: The skill handles untrusted external data via image URLs and user prompts, which represents a potential attack surface.
- Ingestion points: The
images(array of URLs) andprompt(string) fields defined in the input schema in SKILL.md. - Boundary markers: The documentation in SKILL.md notes that the CLI passes prompts as JSON strings to avoid shell expansion risks.
- Capability inventory: The skill utilizes the
runcomfyCLI to perform network operations tomodel-api.runcomfy.netand write results to the local file system. - Sanitization: The skill does not provide specific instructions for validating or sanitizing the content of external image URLs or input prompts.
Audit Metadata