rde-eval

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

• Third-party content exposure detected (high risk: 0.75). The required workflow includes cloning and reading outsider-authored repository files at runtime (e.g., git:https://github.com/... specs in cloud mode), which then become LLM-readable text for analysis, creating an indirect prompt-injection risk.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill uses the git spec git:https://github.com/millionco/react-doctor@ (e.g. git:https://github.com/millionco/react-doctor@main) at runtime to clone and build the monorepo in sandbox workers, which fetches remote code that is then executed in the worker environment.