building-ai-agent-on-cloudflare

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). Documentation includes examples that transmit environment secrets to external servers and expose full agent state/debug endpoints — patterns that can be abused for credential theft and data exfiltration.

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

• Third-party content exposure detected (high risk: 0.85). Yes—runtime WebSocket/HTTP client messages (outsider-authored chat content) are parsed in onMessage and appended into messages/this.state.messages, which are then sent as the messages context to this.env.AI.run(...) (indirect prompt injection via user-provided text).