# Agent Auditor Skill

<figure><img src="/files/t05fxWPGfGtrTLSR6WZ9" alt=""><figcaption></figcaption></figure>

The **Auditor Skill** enables administrators to audit AI conversations for **security, privacy, and compliance** purposes. It automatically scans messages for:

* Personally Identifiable Information (PII)
* Hacking, system abuse, or malicious intent
* Custom organization-defined compliance rules

Detected alerts are **normalized, logged**, and can be **sent to external monitoring systems via webhook**. Admins can also define custom prompts to ensure conversations remain aligned with internal policies.

### 🛡️ Configuration Overview

<figure><img src="/files/kUHUu8un71SirHwZArJx" alt=""><figcaption></figcaption></figure>

The Auditor Skill configuration is broken into two parts:

1. **Auditors** — which audit systems are active
2. **Recipients** — who gets notified when alerts occur

***

### 🕵️ PII Auditor (Optional)

**Purpose:** Scans messages for personal identifiers such as emails, phone numbers, and sensitive user data.

**How to Enable:**

* Toggle **Enable PII Auditor**

> Recommended as a core safeguard against data leakage.

***

### 💻 Hack Auditor (Optional)

**Purpose:** Detects hacking patterns, exploitation attempts, and abusive behaviors.

**How to Enable:**

* Toggle **Enable Hack Auditor**

> Helps protect the system from misuse or malicious intent.

***

### ✍️ Custom Auditor (Optional)

**Purpose:** Enforce organization-specific policies using custom rules, keywords, or prompts.

**How to Enable:**

* Toggle **Enable Custom Auditor**
* Configure custom audit criteria

> Ideal for regulatory, compliance, or industry-specific monitoring.

***

### 📬 Recipients (Required)

**Purpose:** Defines where auditor alerts are delivered.

**How to Configure:**

1. Select a **Channel** (e.g., Webhook)
2. Enter the **Recipient Endpoint** (URL, email, etc.)
3. Optionally **+ Add Header** for authorization or metadata
4. Click **+ Add Recipient** to send alerts to multiple destinations

✅ *At least one recipient must be configured for alerts to be dispatched.*

***

### 📊 Alerts View

<figure><img src="/files/AGViHOvx3u4jLZzNs5OG" alt=""><figcaption></figcaption></figure>

The **Alerts** tab provides centralized monitoring of triggered audit events.

Features:

* **Search & Filter** by:
  * Auditor type (PII, Hack, Custom)
  * Date range
* **Alerts List**
  * Displays alerts matching current filters
  * Shows: *“No auditor alerts match the current filters”* if none found

***

### ✅ Best Practices

* **Start with PII Auditor**\
  Default safeguard for user data protection
* **Use Webhooks for Automation**\
  Easily integrate with Slack, Splunk, SIEM tools, etc.
* **Grow Custom Rules Over Time**\
  Begin with high-priority compliance requirements
* **Review Alerts Regularly**\
  Refine filters and accuracy as patterns emerge


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.raiaai.com/raia-cx/agent-skills/agent-auditor-skill.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.