gpt-image-edit
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the local
runcomfyCLI to perform image processing tasks. Instructions guide the user to invokeruncomfy run openai/gpt-image-2/editwith JSON-formatted inputs. - [EXTERNAL_DOWNLOADS]: The skill requires the installation of the
@runcomfy/clipackage from the npm registry and provides an installation command for a skill collection from theagentspace-soGitHub repository. - [CREDENTIALS_UNSAFE]: While the skill manages an API token (
RUNCOMFY_TOKEN), it provides explicit guidance on secure storage (mode 0600) and recommends the use of environment variables for non-interactive environments like containers or CI/CD pipelines. - [DATA_EXFILTRATION]: Network operations are restricted to the official service domains (
runcomfy.comandruncomfy.net) for the purpose of submitting requests and downloading generated image assets. - [PROMPT_INJECTION]: The skill provides documentation on prompting patterns to optimize model performance but does not include instructions that attempt to bypass safety filters or override agent behavior.
Audit Metadata