Skip to main content
Version: 6.2

Installing the Content Module

Table of Contents​

Overview​

The AI Security content module is distributed as the ai_security.zip archive, which contains ready-to-use knowledge objects. Import the archive through Content Management.

Module Contents​

Object TypeObjects
Dashboards2 dashboards covering LLM threats, agent security, and model theft
Scheduled jobs32 jobs for threat detection, event processing, and lookup population
Runtime indexesgen_ai_events, gen_ai_gpu_metrics, gen_ai_permissions
Internal indexesStore the results of static and semantic request analysis
TagsAI Security, LLM01, LLM03, LLM04, LLM05, LLM06, LLM07, LLM08, LLM10
ISM policiesLifecycle policy for gen_ai_* indexes
Important

The settings in the indexes and policiesism directories must be adapted to the data retention and rotation requirements of the target deployment.

Installation Procedure​

Step 1. Install the Module​

Before importing the archive, install the utility. It prepares the module dependencies, lookups, and semantic search model.

Follow the instructions in Content Management.

Example​

  1. Go to Content Management (Main Menu - System Parameters - Module Settings - General - Content Management)
  2. Select or drag the ai_security.zip archive updated by the utility
  3. Confirm the installation. When the process is complete, the objects become available in the corresponding interface sections

Additional AI Analysis​

You can connect an external AI model to perform additional checks on detected matches:

  1. Disable AI Security: LLM01 — Prompt injection
  2. Configure a connection to the external AI service using the platform tools. In the AI Security: LLM01 — Prompt injection AI check query, replace __EXTERNAL_AI_MODEL__ with the identifier of an available model
  3. Enable the AI check job
Important

Only one of the LLM01 incident creation jobs should be enabled at a time: Prompt injection or Prompt injection AI check.

Step 2. Verify the Imported Module Objects​

1. Verify the Dashboards​

Verify that the two module dashboards are available:

Main Menu - Main - Dashboards

Dashboards

2. Verify the Scheduled Jobs​

Verify that the 32 module jobs are available:

Main Menu - Job Scheduler - Jobs List

Jobs

3. Verify the Indexes​

Verify that the indexes are available:

Main Menu - System Parameters - Index Management - Indexes

Indexes

4. Verify the Index Management Policy​

Verify that the ISM policy is available:

Main Menu - System Parameters - Index Management - Index Policies

ISM policies

5. Verify the Tags​

Verify that the tags are available:

Main Menu - System Parameters - Module Settings - Tags

  • AI Security
  • LLM01: Prompt Injection
  • LLM03: Training Data Poisoning
  • LLM04: Model Denial of Service (DoS)
  • LLM05: Supply Chain Vulnerabilities
  • LLM06: Sensitive Information Disclosure
  • LLM07: Insecure Plugin Design
  • LLM08: Excessive Agency
  • LLM10: Model Theft