How it works
The Auditor skill acts as a security and compliance firewall for your agent. It monitors all inbound and outbound messages in real-time, detecting sensitive data, malicious inputs, or policy violations before they are processed.
Key features
PII Detection
Automatically redact or block PII (SSN, credit cards).
Prompt Injection Defense
Detect and block attempts to manipulate the agent's instructions (jailbreaks).
Custom Policies
Define specific topics the agent is strictly forbidden from discussing.
Alerting
Trigger alerts to administrators when a violation occurs.
How to configure it
- 01
Enable the Auditor skill in Launch Pad.
- 02
Choose which auditor modules to enable (PII, Hack/Injection, Custom).
- 03
Decide whether sensitive data should be blocked entirely or redacted (e.g., replaced with [REDACTED]).
- 04
Enter the email addresses or webhook URLs that should receive notifications when a violation is detected.
Frequently asked questions
Does the Auditor increase response time?
Slightly — because it evaluates messages before they reach the main LLM, it adds a small amount of latency to ensure compliance.
Can I define custom regex patterns for PII?
Yes — the Custom Auditor module allows you to define regular expressions for proprietary data formats unique to your business.