WARNING！！！LOTS OF VIBE CODING！！

AJS is a specialized tool designed to assist security researchers and developers in understanding and reverse-engineering complex, obfuscated front-end JavaScript, with a primary focus on cryptographic functions.

It leverages the Chrome DevTools Protocol (CDP) and Large Language Models (LLMs) to turn the tedious process of manual debugging and code analysis into a fast, AI-driven query.

The tool's architecture is simple and powerful:

1. Live Debugging: AJS connects to your running Chrome browser instance. Its CDPDebugger component sets dynamic breakpoints on key user interaction events (click, submit, change).

2. Contextual Snippet Extraction: When you perform an action on a web page (e.g., clicking a "Login" button that triggers encryption), the debugger catches the event. Instead of dumping the entire script, it intelligently extracts a small, relevant snippet of the JavaScript code around the breakpoint, along with the values of local variables at that exact moment.

3. AI-Powered Analysis: This high-value "intelligence packet" (code snippet + variables) is sent to an LLM. The LLM is given a specific persona: a JavaScript Reverse Engineering Expert. It is tasked with:

   • Identifying the encryption algorithm (AES, RSA, etc.).
   • Locating cryptographic keys, IVs, and other constants in the code or variables.
   • Providing a step-by-step explanation of the encryption logic.

4. Instant Results: The AI's analysis is printed directly to your console, giving you immediate insight into how the front-end encryption works.

This process transforms a potentially hours-long RE task into a simple action-and-response loop.

• Python 3.10+
• A running instance of Google Chrome or a Chromium-based browser.
• Access to an OpenAI-compatible LLM API (e.g., a local LM Studio, Ollama, or the official OpenAI API).

pip install -r requirements.txt

For AJS to connect, you must start Chrome with a remote debugging port enabled. First, close all existing Chrome instances. Then, run the appropriate command for your OS:

Windows

"C:\Program Files\Google\Chrome\Application\chrome.exe" --remote-debugging-port=9222

macOS

/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --remote-debugging-port=9222

Linux

google-chrome --remote-debugging-port=9222

Edit the config.yaml file:

• whitelist_domains: (IMPORTANT) Add the domains you intend to analyze (e.g., example.com, localhost). This is a crucial security measure.
• llm_service.api_config: Configure the base_url, model_name, and api_key for your chosen LLM service.

python main.py

Once running, simply use your main Chrome browser. Navigate to the whitelisted page you want to analyze. Perform an action that you suspect triggers encryption (like clicking a login or submit button). Watch the console where you ran AJS. The AI's reverse engineering analysis will be printed there shortly after the event is triggered.