image-generation

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). Yes — the prompt includes explicit directives to "Do NOT read the python file" and "You don't need to check the folder under /mnt/user-data", which instruct the agent to avoid inspecting code or user data and thus hide potential behavior outside the advertised image-generation purpose.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs (in "Tips: Enhancing Generation with Reference Images" and the example workflow) to call the image_search tool and download returned image URLs to local files as reference images, which the generate.py then reads and embeds in the generation request—meaning untrusted third‑party images from the open web are fetched and can materially influence generation behavior.