image-upscaling

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

• Third-party content exposure detected (high risk: 0.70). This skill's SKILL.md explicitly instructs the agent to fetch and process arbitrary public image URLs (e.g., the examples using "https://your-image.jpg" and the workflow step passing ""), meaning untrusted third-party content from the open web is ingested and could contain instructional content that materially influences downstream processing.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill requires and instructs use of the inference.sh CLI to run remote apps (e.g., "infsh app run ..."), which causes execution of remote code on https://inference.sh at runtime, so that external URL controls execution required by the skill.