The Agent Skills Directory

[SAFE]: The skill relies on official Cloudflare documentation and vendor-specific libraries, including @cloudflare/codemode and @cloudflare/ai-chat.
[SAFE]: Documentation for experimental features like 'Code Mode' explicitly includes security controls, such as a globalOutbound fetch handler to implement domain whitelisting and prevent unauthorized network access.
[SAFE]: Email handling logic demonstrates secure practices by providing examples of HMAC-SHA256 signature verification (createSecureReplyEmailResolver) to ensure the authenticity of inbound messages.
[SAFE]: The skill documentation uses standard, non-obfuscated instructions and follows the recommended platform architecture for Cloudflare Workers and Durable Objects.

agents-sdk