Skip to content

AI Assistant

The AI Assistant provides intelligent chat-based support for analyzing scan data, executing tasks, and answering questions about your reconnaissance activities. The assistant integrates seamlessly with your Reverge workflow and maintains context awareness of your current work.

Overview#

The AI Assistant appears as a resizable chat panel on the right side of the interface, providing real-time AI-powered assistance without interrupting your main workflow. The assistant can help with data analysis, vulnerability identification, scan execution, and general security questions.

AI Assistant Panel


Getting Started#

Accessing the AI Assistant#

The AI assistant panel is available on most pages within Reverge. Look for the chat panel on the right side of your screen. If not visible, it may be collapsed - look for an expand button or toggle.

Selecting an AI Provider#

Before using the assistant, you must select an AI provider from the dropdown at the bottom of the chat panel. The available providers depend on your configured AI models in the settings.

AI Provider Selection

Chat Interface#

Sending Messages#

  • Type your message in the input field at the bottom of the panel
  • Press Enter to send your message
  • Press Shift+Enter to add a new line without sending
  • Click the send button (📤) to send your message

Message Types#

  • User messages appear in blue bubbles on the right side
  • AI responses appear in white bubbles on the left side with the AI avatar
  • Streaming responses show content as it's being generated with a typing cursor

Features#

Context Awareness#

The AI assistant automatically understands your current context: - Scan Data: When viewing scan results, the assistant can analyze findings - Target Information: When on target pages, the assistant knows the target details - Port Analysis: When examining specific ports, the assistant has port context - Vulnerability Data: The assistant can reference current vulnerability findings

Chat Persistence#

  • Server-Side Storage: Chat history is saved to the Reverge database and persists across sessions, devices, and browser restarts
  • Per-User History: All conversations are scoped to your account and only visible to you
  • Session Browser: Access previous conversations at any time using the history panel
  • Automatic Continuation: Returning to the assistant resumes where you left off in your most recent session

Streaming Responses#

  • Real-time Generation: See AI responses as they're being generated
  • Cancellable Requests: Stop generation mid-stream if needed
  • Connection Recovery: Automatic handling of connection issues

Advanced Features#

Auto Mode#

Auto Mode turns the AI assistant into a fully autonomous penetration testing agent. When enabled, a built-in Auto Mode Operator agent drives the session: it analyses the current Reverge data (discovered hosts, open ports, scan results), issues a directive to the Reverge AI agent, evaluates the response, and repeats — forming a continuous operator → agent loop that runs without user intervention.

All scan jobs scheduled during an Auto Mode session are routed through the Reverge scheduler and remain visible in the normal scan history.

Enabling Auto Mode

  1. Open the AI assistant chat panel.
  2. Toggle the Auto Mode switch located below the chat input area.
  3. Type your high-level objective (e.g. "Enumerate all services on target example.com and report any critical findings") and press Enter to send.
  4. The operator takes over from the first AI response and continues issuing directives autonomously.

While Auto Mode is Running

  • The operator bubble (labelled Operator) shows the directive issued for each turn, together with any thinking steps and tool calls the operator made to arrive at it.
  • The agent bubble (labelled REVERGE AI) shows the agent's response to each directive, including intermediate tool-call and thinking step indicators.
  • The token counter (top-right of the input area) tracks cumulative token usage across both the operator and agent for the entire session.
  • A turn indicator below the toggle (e.g. Turn 3/10) shows current progress against any configured limits.

Stopping Auto Mode

Toggle the Auto Mode switch off at any time to stop the loop gracefully after the current agent response completes. The session history is preserved and the chat remains available for normal use.

Auto Mode also stops automatically when any configured limit is reached (max turns, max time, max tokens, or when a significant finding is discovered if Stop on Significant Findings is enabled in Settings → AI → Auto Mode).

Chat History Panel#

  • Browse Sessions: Click the history button (🕐) in the header to open the history sidebar listing all previous conversations
  • Resume a Session: Click any session in the list to load its messages and continue the conversation
  • Search History: Use the search bar in the history panel to run a full-text search across all past messages
  • Delete Sessions: Remove individual sessions from the history panel using the delete button next to each entry

Panel Management#

  • Resizable Panel: Drag the left edge to resize the chat panel horizontally

  • Size Constraints: Panel maintains minimum (300px) and maximum (800px) widths

  • Auto-height: Panel automatically adjusts to your browser window height

Markdown Support#

The AI assistant supports rich text formatting:

  • Code Blocks: Properly formatted code snippets with syntax highlighting

  • Lists: Bulleted and numbered lists for organized information

  • Emphasis: Bold and italic text formatting

  • Headers: Structured content with headers and subheaders

Use Cases#

Scan Analysis#

Ask the assistant to help interpret scan results:

  • "What vulnerabilities were found in the latest scan?"

  • "Explain the significance of open port 445"

  • "What's the risk level of the discovered services?"

Task Execution#

Request help with specific tasks:

  • "Schedule a scan for this target"

  • "How do I configure a custom scan profile?"

  • "Show me recent scan history"

Vulnerability Research#

Get detailed information about security issues:

  • "Tell me about CVE-2023-1234"

  • "What are the best practices for securing Apache servers?"

  • "How critical is this SQL injection finding?"

Data Queries#

Ask questions about your reconnaissance data:

  • "How many hosts were discovered this week?"

  • "What domains are associated with this target?"

  • "Show me all findings for port 80"

Chat Controls#

Header Controls#

  • New Chat (➕): Start a fresh conversation (previous sessions remain in history)
  • History (🕐): Open the chat history panel to browse, search, and resume past conversations
  • Clear Chat (🧹): Clear the current conversation view while keeping it in history
  • Settings (⚙️): Quick access to AI model configuration settings

Input Controls#

  • Send Button: Changes to a stop button during AI generation

  • Provider Selection: Choose between configured AI providers

  • Auto-resize Input: Text area expands as you type longer messages

  • Auto Mode Toggle: Enable or disable the autonomous operator → agent loop for the current session. The toggle is located below the input field alongside the session token counter.

Troubleshooting#

No AI Providers Available#

If you see "No providers configured":

  1. Navigate to Settings > AI Models

  2. Configure at least one AI provider

  3. Return to refresh the AI assistant

Connection Issues#

If responses aren't generating:

  • Check your internet connection

  • Verify AI provider API credentials in settings

  • Try selecting a different AI provider

  • Refresh the page and try again

Slow Responses#

  • Some AI providers may have slower response times

  • Try switching to a faster provider if available

  • Check for rate limiting in your AI provider account

Chat History Lost#

  • Chat history is stored server-side and tied to your user account

  • History persists across browser sessions, incognito windows, and devices

  • Individual sessions can be deleted from the history panel

  • Contact your administrator if chat history is unexpectedly missing

Tips and Best Practices#

Effective Prompting#

  • Be Specific: Include relevant details about your current task
  • Ask Follow-ups: Build on previous responses for deeper analysis
  • Use Context: Reference specific scan results or targets when applicable

Performance Optimization#

  • Clear Old Chats: Regularly clear chat history for better performance
  • Choose Appropriate Providers: Use faster providers for simple queries
  • Cancel Unneeded Requests: Stop generation if the response isn't helpful

Privacy Considerations#

  • Sensitive Data: Be mindful of what information you share with AI providers
  • API Costs: Some providers charge per token - monitor usage
  • Data Retention: Check your AI provider's data retention policies

Security Note: Be cautious when sharing sensitive reconnaissance data with external AI providers. Consider using local models (Ollama) for highly sensitive information.