🛡️ Prompt Injection Protection
Configure security wrappers for untrusted external inputs. Prevents prompt injection attacks on AI agents.
🔍 Channel Trust Levels
Each channel has a trust level based on configured rules. Untrusted channels get security wrappers automatically.
| Channel Pattern | Trust Level | Protection | Status |
|---|---|---|---|
| Loading... | |||
🛡️ Sanitization Rules
Rules that wrap untrusted input with security context before agent processing.
| On | Channel | Priority | Description | Template Preview | Actions |
|---|---|---|---|---|---|
| Loading... | |||||
🧪 Test Sanitization
Test how untrusted input would be wrapped by your rules.
📊 Sanitization Statistics
—
Active Rules
—
Protected Channels
—
Inputs Sanitized (24h)