illustrated-explainer-spec
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implementation includes a web server that accepts user-supplied query strings and interpolates them directly into prompts for an AI image generation model.
- Ingestion points: The
queryfield in the/api/pageendpoint inserver.jsaccepts user input. - Boundary markers: The prompt templates in
server.jsdo not use explicit delimiters or instructions to ignore embedded commands within the user-provided data. - Capability inventory: The server implementation is capable of making network requests to AI model providers (Google Gemini or OpenAI) and writing generated image files to the local file system via the
fsmodule inserver.js. - Sanitization: The input is validated in
server.jsto ensure it is a string between 1 and 300 characters and is trimmed of leading/trailing whitespace.
Audit Metadata