background-removal

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow uses belt app run ... --input with image_url values that point to arbitrary external URLs; the CLI will fetch and ingest the referenced page/image content as readable text/metadata into the LLM context, creating an outsider-supplied content path (public web content).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes remote apps via the inference.sh CLI (e.g., "belt app run infsh/birefnet" and "belt app run falai/reve"), which executes code on the https://inference.sh platform at runtime and is required for the skill to function.