design-review

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

• Third-party content exposure detected (low risk: 0.10). The skill performs web browsing and searches arbitrary target URLs provided by the user or dynamically, exposing the agent to outsider-authored free text via rendered web content.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill executes setup scripts and retrieves installer content from https://bun.sh/install at runtime, which constitutes an operational dependency with weak/unknown provenance because the remote script execution lacks enforced cryptographic integrity or strong publisher guarantees.