Browser Extensions
Browser extension settings
Configure personalization, privacy scans, prompt protection and action safety.
| Setting | Effect |
|---|---|
| Governed GitHub transactions | Turns eligible GitHub mutations Off, Observe or Enforce; server policy remains authoritative |
| Microphone access | Enables voice-to-text input after browser permission is granted |
| Personalized responses | Allows memory-based personalization; requires Premium or higher |
| Light Prompt Guardrails | Fast checks for prompt injection and security-sensitive requests |
| Advanced Prompt Protection | Scans the message, selected page context and selected text attachments where eligible |
| Max Prompt Protection | Extends deep scanning to full selected page context and larger selected attachments |
| Quick Scan | Detects credentials and other sensitive data in messages, context and attachments |
| Deep LLM Scan — Chats | Uses deeper scanning for personal information in chat messages |
| Deep LLM Scan — Files | Uses deeper scanning for attachments and context files |
| Agent action-safety details | Shows policy decisions without weakening enforced policy |
| Style Profile | Applies selected personal conventions when Cortex generates code |
Quick Scan is disabled while Deep LLM Scan for chats is enabled. Deep and Max protection can require an eligible plan, model download and additional processing time. Organization policy can enforce stricter behavior than the local setting.

