gen-image
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Node.js script
scripts/image.mjsto perform image generation and editing tasks. - [EXTERNAL_DOWNLOADS]: The script communicates with Google's official Generative Language API (
generativelanguage.googleapis.com) to process image requests. This is a well-known service and the communication is required for the skill's primary purpose. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes user-provided prompts and input images to generate content via an LLM/Image model.
- Ingestion points: User-provided strings in the
promptargument and content from local image files passed toscripts/image.mjs. - Boundary markers: None identified; user input is directly interpolated into the API request body.
- Capability inventory: The script can read local files (
fs.readFileSync), create directories (fs.mkdirSync), write local files (fs.writeFileSync), and perform network requests (fetch). - Sanitization: No input sanitization or validation of the prompt content is implemented before it is sent to the remote API.
Audit Metadata