cloudflare-email-service

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

• Third-party content exposure detected (medium risk: 0.30). In the required runtime flow for receiving/routing email, the Worker email() handler ingests outsider-authored inbound email content (e.g., message.raw, headers/subject/body) that was authored by the sender and is then parsed/processed at runtime.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to "prefer retrieval" and to fetch authoritative docs at runtime (e.g., https://github.com/cloudflare/agents/tree/main/docs and https://developers.cloudflare.com/api/resources/email_sending), so these external URLs can be fetched during runtime and their content injected into the agent's context, directly controlling prompts/instructions.