skills/jau123/meigen-art/AI Image Generation & Editor — Nanobanana, GPT Image, ComfyUI/Gen Agent Trust Hub
AI Image Generation & Editor — Nanobanana, GPT Image, ComfyUI
Pass
Audited by Gen Agent Trust Hub on Mar 24, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill executes the
meigenpackage at version 1.2.5 from the NPM registry vianpx. This is the core component for providing image generation tools. - [EXTERNAL_DOWNLOADS]: Interacts with the vendor's API and other configured AI provider endpoints to fetch model data and generated image files.
- [DATA_EXFILTRATION]: Provides a mechanism for users to upload local images to the MeiGen cloud service to be used as visual references. This feature is clearly documented as a requirement for certain generation modes.
- [COMMAND_EXECUTION]: Leverages
mcporterandnpxcommands for tool orchestration and to start the MCP server process. - [CREDENTIALS_UNSAFE]: Instructs users to manage API keys and tokens within local environment variables or specific configuration files (~/.config/meigen/config.json), which is consistent with secure local development practices.
- [PROMPT_INJECTION]: The skill processes untrusted user input and gallery search results (ingestion points in SKILL.md) through prompt enhancement and generation tools. While specific boundary markers are absent in the prompt interpolation, the instructions include strict rules to limit agent commentary and ensure tool responses are presented literally, which mitigates the risk of indirect prompt injection.
Audit Metadata