AI Assistant

Overview

The AI Assistant uses YandexGPT to provide intelligent help with CIS benchmarks, compliance scanning, and hardening troubleshooting.

Navigation

Menu: AI Tools > AI Settings

Features

• Chat Interface - Ask questions about security compliance
• LLM Configurations - Manage YandexGPT settings
• Rule Explanations - AI explains compliance scan rule failures
• Usage Statistics - Track AI usage and costs

Page Layout

The AI Settings page has three tabs:

Chat Tab

Interactive chat interface with the AI assistant.

Configs Tab

Manage LLM configurations.

Usage Tab

View usage statistics with three statistic cards:

• Total Requests - Number of AI API calls
• Input Tokens - Total tokens sent to AI
• Output Tokens - Total tokens received from AI

Configuration Management

Configs Table Columns

Adding Configuration

1. Click Add Config button
2. Fill in the configuration form:

*Either API Key or IAM Token is required.

3. Click Create

Syncing Models

The Model dropdown shows available models from Yandex Cloud.

1. Click Sync Models button next to the Model label
2. Wait for sync to complete
3. Select from updated model list

If models need sync, a helper text appears below the dropdown.

Editing Configuration

1. Click Edit button on a configuration row
2. Modify fields in the modal
3. Click Update

Testing Configuration

1. Click Test button on a configuration row
2. Wait for test to complete
3. Success shows:
   • Success message with response time
   • Response text
   • Input/Output tokens
   • Duration in milliseconds
4. Failure shows error message

Deleting Configuration

1. Click Delete button on a configuration row
2. Confirm deletion in the popup

Chat Interface

Starting a Conversation

1. Go to AI Settings page
2. Chat tab is selected by default
3. Optionally select a specific configuration from the dropdown
4. Type your question in the input field
5. Press Enter or click Send

If no configuration is selected, the default active configuration is used.

Chat Features

• Markdown Support: AI responses render as formatted markdown
• User Messages: Displayed on the right with blue background
• AI Messages: Displayed on the left with robot icon
• Loading Indicator: Shows “Thinking…” while waiting for response
• Auto-scroll: Chat scrolls to show new messages

Welcome Screen

When no messages exist, the chat shows:

• Robot icon
• Welcome message
• List of example topics the AI can help with

Example Questions

• “What is CIS benchmark rule 1.1.1.1?”
• “How do I fix SSH configuration issues?”
• “What profile should I use for a web server?”
• “Explain the difference between Level 1 and Level 2”
• “Why might hardening disable my application?”

Chat Capabilities

The assistant can help with:

• CIS benchmark rules and explanations
• Choosing appropriate compliance profiles
• Understanding scan results
• Troubleshooting hardening failures
• Security configuration recommendations
• Best practices for server hardening

Clearing Chat

Click Clear button to start a new conversation.

Keyboard Shortcuts

• Enter: Send message
• Shift+Enter: New line in message

AI Features in Other Pages

Rule Explanation (Compliance Page)

When LLM is configured, an Explain button appears in compliance scan results:

1. Go to Compliance page
2. View scan results
3. Click Explain button on a rule row, or
4. Expand the rule and click Explain with AI
5. A modal shows AI-generated explanation:
   • Rule ID, Severity, and Result tags
   • Rule title
   • Detailed explanation in markdown format

Model Selection Guide

Best Practices

Effective Questions

• Be specific about your situation
• Include relevant error messages
• Mention the OS and profile used
• Ask one question at a time

Configuration

• Use lower temperature (0.3) for factual answers
• Use higher temperature (0.7) for creative suggestions
• Set reasonable max_tokens to control response length
• Keep one configuration active and set as default

Cost Control

• Disable unused configurations
• Monitor usage statistics on the Usage tab
• Use yandexgpt-lite model when possible
• Clear chat to reduce context length

Troubleshooting

No Response from AI

1. Verify configuration is active
2. Test configuration connection with Test button
3. Check API key is valid and not expired
4. Verify Catalog ID is correct

“No active configuration” Warning

1. Go to Configs tab
2. Ensure at least one configuration is active (green status)
3. Toggle the Active switch on a configuration

Slow Responses

1. Try yandexgpt-lite model
2. Reduce max_tokens setting
3. Clear chat history
4. Check network connection

Invalid Responses

1. Rephrase question
2. Provide more context
3. Lower temperature for factual queries
4. Try different model

Configuration Test Fails

1. Verify API key is correct
2. Check Catalog ID matches your account
3. Ensure YandexGPT is enabled in your cloud
4. Check IAM permissions (ai.languageModels.user role)
5. Try syncing models to verify connectivity

Models Not Loading

1. Click Sync Models button
2. Check for network errors
3. Verify API key has correct permissions

Yandex Cloud Setup

Getting API Key

1. Go to Yandex Cloud Console
2. Navigate to IAM > API Keys
3. Create a new API key
4. Copy the secret key value

Getting Catalog ID

1. Go to Yandex Cloud Console
2. Navigate to your folder
3. Copy the Folder ID (this is your Catalog ID)

Required Permissions

The API key needs the ai.languageModels.user role to access YandexGPT.

Related Pages

• Compliance Scanning - AI rule explanations in scan results
• Settings - General platform settings