import-prototype

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

• Third-party content exposure detected (high risk: 0.75). The required workflow takes a user-supplied Builder.io prototype URL and runs npx @builder.io/dev-tools ... code --prototype "{builderUrl}", which at runtime will fetch the prototype’s content from Builder.io (outsider-authored free text) and feed it into the agent/CLI process that can be included in LLM context via tool output/logs.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill runs "npx --yes @builder.io/dev-tools@latest" at runtime, which will fetch and execute remote npm package code (a required runtime dependency), so remote code is executed during skill operation.