design-lab

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow is to install and invoke the design-lab skill from the public GitHub package 0xdesign/design-plugin; at runtime the agent will fetch and execute that outsider-authored skill code/content, which can ingest free-form text from external sources during the design interview/variant generation flow.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill stub requires fetching and installing the external skill code at runtime (e.g., via "npx skills add 0xdesign/design-plugin --skill design-lab" which pulls from https://github.com/0xdesign/design-plugin), and that remote package will supply the skill logic and prompts that control agent behavior, so this is a runtime external dependency that can execute/control instructions.