cloudflare-worker-dev

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

• Third-party content exposure detected (high risk: 0.80). This skill includes examples that fetch and proxy arbitrary external URLs (e.g., "const response = await fetch(upstreamUrl, request)" in Quick Reference and the anti-pattern "fetch('https://slow-api.com/data')") and also accepts user-generated WebSocket messages in the Durable Objects chat room, so it clearly ingests untrusted third-party content as part of its runtime.