Learn how to control which websites your AI agent can visit. This guide covers allowlists, blocklists, prompt injection risks, and code-level guardrails — so your agent browses safely and only goes where you intend.