Interactive Mode

By default, Strix runs in interactive mode with a text-based user interface (TUI) that provides real-time visibility into the penetration testing process.

Overview

Interactive mode is the default when you run Strix without the --non-interactive flag:

strix --target https://example.com

The TUI provides:

Real-time agent activity - Watch AI agents work in real-time
Vulnerability notifications - See vulnerabilities as they’re discovered
Agent tree view - Track parent and child agents
Live statistics - Monitor progress and resource usage
Interactive chat - Guide agents during the scan

TUI Interface

Layout

The TUI is divided into several sections:

┌─────────────────────────────────────────────────────┬──────────────────┐
│                                                     │                  │
│                                                     │   Agents Tree    │
│                                                     │                  │
│              Chat / Agent Activity                  ├──────────────────┤
│                                                     │                  │
│                                                     │ Vulnerabilities  │
│                                                     │                  │
│                                                     ├──────────────────┤
│                                                     │                  │
│                                                     │   Statistics     │
├─────────────────────────────────────────────────────┤                  │
│ > [Chat Input]                                      │                  │
└─────────────────────────────────────────────────────┴──────────────────┘

The left sidebar shows all active agents in a tree structure:

⚪ Running - Agent is actively working
⏸ Waiting - Agent is waiting for input or another agent
🟢 Completed - Agent finished successfully
🔴 Failed - Agent encountered an error
■ Stopped - Agent was manually stopped

Agents with vulnerabilities show a count: ⚪ Web Scanner (3) Vulnerabilities are displayed with severity-colored indicators:

● Critical - Red
● High - Orange
● Medium - Amber
● Low - Green
● Info - Blue

Click on any vulnerability to view full details.

Chat Area (Main Panel)

The main panel shows:

Agent messages - LLM reasoning and decision-making
Tool executions - What tools agents are using
Streaming output - Real-time content as it’s generated
Status indicators - Current agent state

Displays:

Scan duration
Active agents count
Tools executed
Vulnerabilities found
Version number

Keyboard Shortcuts

Enter

key

Send a message to the selected agent

Shift+Enter

key

Insert a newline in the chat input

Tab

key

Switch between chat input and agents tree

↑ / ↓

key

Navigate the agents tree

Esc

key

Stop the selected agent

key

Show help screen

Ctrl+Q / Ctrl+C

key

Quit Strix (with confirmation)

Interacting with Agents

Selecting an Agent

Press Tab to focus the agents tree
Use ↑ and ↓ arrows to select an agent
Press Tab again to return to chat input

The main panel updates to show the selected agent’s activity.

Sending Messages

You can guide agents by sending messages:

Select an agent from the tree
Type your message in the chat input
Press Enter to send

Example messages:

Focus on the authentication endpoints

Try using SQL injection on the search parameter

Check for IDOR vulnerabilities in /api/users

Resuming Waiting Agents

When an agent status shows ⏸ Waiting, it needs your input:

Select the waiting agent
Read its last message to understand what it needs
Send a response to resume the agent

Stopping Agents

Select the agent you want to stop
Press Esc
Confirm the action when prompted

Viewing Vulnerabilities

In the Vulnerabilities Panel

Click on any vulnerability in the sidebar
A detailed modal opens showing:
- Full description
- CVSS score and breakdown
- Technical analysis
- Proof-of-concept code
- Remediation steps

Copying Vulnerability Reports

In the vulnerability detail modal:

Click the Copy button
The vulnerability is copied to clipboard in Markdown format
Paste into GitHub issues, Jira, or documentation

Startup Sequence

1. Splash Screen

When you launch Strix, you’ll see an animated splash screen:

 ███████╗████████╗██████╗ ██╗██╗  ██╗
 ██╔════╝╚══██╔══╝██╔══██╗██║╚██╗██╔╝
 ███████╗   ██║   ██████╔╝██║ ╚███╔╝
 ╚════██║   ██║   ██╔══██╗██║ ██╔██╗
 ███████║   ██║   ██║  ██║██║██╔╝ ██╗
 ╚══════╝   ╚═╝   ╚═╝  ╚═╝╚═╝╚═╝  ╚═╝

Welcome to Strix!
Starting Strix Agent...

This displays for approximately 4.5 seconds while:

Environment is validated
LLM connection is tested
Docker image is pulled (first run only)
Agent configuration is prepared

2. Main Interface

After the splash screen, the full TUI interface loads and the scan begins automatically.

Scan Completion

When the scan completes:

In Interactive Mode

The TUI remains open so you can:

Review the full scan history
Examine vulnerabilities in detail
Copy reports to clipboard
Navigate the agent tree

Press Ctrl+Q to quit when you’re ready.

Completion Panel

When you quit, you’ll see a summary panel:

╔══════════════════════════════════════════╗
║ STRIX                                    ║
║ Penetration test completed               ║
║                                          ║
║ Target  https://example.com              ║
║                                          ║
║ Vulnerabilities  3 found                 ║
║ Duration         12m 34s                 ║
║ Tools executed   47                      ║
║                                          ║
║ Output  strix_runs/example_20260301_1234║
╚══════════════════════════════════════════╝

models.strix.ai  ·  discord.gg/strix-ai

Terminal Requirements

Minimum Requirements

Terminal emulator with 256 color support
Minimum size: 80x24 characters
Recommended size: 120x40 characters or larger

Supported Terminals

Linux: GNOME Terminal, Konsole, Alacritty, kitty, Terminator
macOS: Terminal.app, iTerm2, Alacritty, kitty
Windows: Windows Terminal, WSL2 with any Linux terminal

Font Recommendations

For best results, use a monospace font that supports Unicode box-drawing characters:

Fira Code
JetBrains Mono
Cascadia Code
SF Mono (macOS)
Consolas (Windows)

Advanced Features

Multi-Agent Coordination

The agents tree shows parent-child relationships:

⚪ Strix Agent
  ├─ ⚪ Web Scanner (1)
  │   ├─ 🟢 Auth Tester
  │   └─ ⚪ API Fuzzer (2)
  └─ ⚪ Code Analyzer
      └─ 🟢 SAST Agent

This visualizes:

Which agents spawned which sub-agents
Agent hierarchy and specialization
Which agents found vulnerabilities

Real-Time Streaming

As agents think and work, you see:

Text segments - Agent reasoning in natural language
Tool calls - Functions being invoked with parameters
Completion markers - When tools finish executing

Live Statistics Updates

Statistics update every 350ms showing:

How long the scan has been running
How many agents are active
Total tool executions
Vulnerabilities discovered so far

Troubleshooting

TUI Not Displaying Correctly

Problem: Characters are garbled or boxes don’t align Solutions:

Ensure your terminal supports 256 colors: echo $TERM
Try setting: export TERM=xterm-256color
Use a different terminal emulator
Increase terminal font size

Agent Stuck on “Initializing”

Problem: Agent shows status “Initializing” for a long time Solutions:

Wait - LLM initialization can take 30-60 seconds
Check network connectivity to LLM provider
Verify LLM API key is valid
Check if Docker containers are running: docker ps

Can’t See Full Vulnerability Details

Problem: Vulnerability text is truncated Solutions:

Increase terminal window size
Click vulnerability to open detail modal
Use the Copy button to get full report
Check output files in strix_runs/ directory

High CPU Usage

Problem: Terminal is consuming significant CPU Explanation: The TUI refreshes frequently (2 times per second) to show real-time updates. This is normal during active scans. Solutions:

Reduce terminal window size
Use --non-interactive mode instead
Close other terminal tabs/windows

Comparison with Non-Interactive Mode

Feature	Interactive Mode	Non-Interactive Mode
TUI Display	Yes	No
Real-time updates	Yes	Limited
Agent interaction	Yes	No
Auto-exit	No	Yes
Exit code 2 for vulns	No	Yes
CI/CD friendly	No	Yes
Human supervision	Yes	No
Resource usage	Higher	Lower

See Non-Interactive Mode for automation use cases.

Commands

Modes

Output

​Overview

​TUI Interface

​Layout

​Agents Tree (Sidebar)

​Vulnerabilities Panel (Sidebar)

​Chat Area (Main Panel)

​Statistics Panel (Sidebar Bottom)

​Keyboard Shortcuts

​Interacting with Agents

​Selecting an Agent

​Sending Messages

​Resuming Waiting Agents

​Stopping Agents

​Viewing Vulnerabilities

​In the Vulnerabilities Panel

​Copying Vulnerability Reports

​Startup Sequence

​1. Splash Screen

​2. Main Interface

​Scan Completion

​In Interactive Mode

​Completion Panel

​Terminal Requirements

​Minimum Requirements

​Supported Terminals

​Font Recommendations

​Advanced Features

​Multi-Agent Coordination

​Real-Time Streaming

​Live Statistics Updates

​Troubleshooting

​TUI Not Displaying Correctly

​Agent Stuck on “Initializing”

​Can’t See Full Vulnerability Details

​High CPU Usage

​Comparison with Non-Interactive Mode

​See Also

Build docs developers (and LLMs) love