- Explore MCP Servers
- AuditLuma
Auditluma
What is Auditluma
AuditLuma is an AI-powered intelligent code auditing system that leverages multiple AI agents and advanced technologies, including Multi-Agent Cooperation Protocol (MCP) and Self-RAG (Retrieval-Augmented Generation), to provide comprehensive security analysis for codebases.
Use cases
Use cases for AuditLuma include auditing large codebases for vulnerabilities, generating security reports for compliance, assisting developers in identifying and fixing security issues, and providing insights into code dependencies and potential risks.
How to use
To use AuditLuma, clone the repository and install the dependencies. Run the command python main.py -d ./goalfile -o ./reports to analyze a target project directory and output the report to the specified directory.
Key features
Key features of AuditLuma include a multi-agent architecture for specialized auditing tasks, MCP for enhanced coordination among agents, Self-RAG technology for improved context understanding, comprehensive security analysis with vulnerability detection and remediation suggestions, visualization capabilities for generating dependency graphs and detailed reports, multi-LLM vendor support, automatic vendor detection, and asynchronous parallel processing for improved performance.
Where to use
AuditLuma can be used in various fields including software development, cybersecurity, and code quality assurance, where secure and reliable code is essential.
Clients Supporting MCP
The following are the main client software that supports the Model Context Protocol. Click the link to visit the official website for more information.
Overview
What is Auditluma
AuditLuma is an AI-powered intelligent code auditing system that leverages multiple AI agents and advanced technologies, including Multi-Agent Cooperation Protocol (MCP) and Self-RAG (Retrieval-Augmented Generation), to provide comprehensive security analysis for codebases.
Use cases
Use cases for AuditLuma include auditing large codebases for vulnerabilities, generating security reports for compliance, assisting developers in identifying and fixing security issues, and providing insights into code dependencies and potential risks.
How to use
To use AuditLuma, clone the repository and install the dependencies. Run the command python main.py -d ./goalfile -o ./reports to analyze a target project directory and output the report to the specified directory.
Key features
Key features of AuditLuma include a multi-agent architecture for specialized auditing tasks, MCP for enhanced coordination among agents, Self-RAG technology for improved context understanding, comprehensive security analysis with vulnerability detection and remediation suggestions, visualization capabilities for generating dependency graphs and detailed reports, multi-LLM vendor support, automatic vendor detection, and asynchronous parallel processing for improved performance.
Where to use
AuditLuma can be used in various fields including software development, cybersecurity, and code quality assurance, where secure and reliable code is essential.
Clients Supporting MCP
The following are the main client software that supports the Model Context Protocol. Click the link to visit the official website for more information.
Content
AuditLuma - 高级代码审计AI系统 🔍
AuditLuma是一个智能代码审计系统,它利用多个AI代理和先进的技术,包括多代理合作协议(MCP)和Self-RAG(检索增强生成),为代码库提供全面的安全分析。
✨ 特性
- 🤖 多代理架构 - 针对不同审计任务的专门代理
- 🔄 MCP(多代理合作协议) - 增强代理之间的协调与合作
- 🔍 Self-RAG技术 - 提高上下文理解和知识检索能力
- 🛡️ 综合安全分析 - 全面检测漏洞并提出有效修复建议
- 📊 可视化功能 - 生成依赖关系图和详细安全报告
- 🌐 多LLM厂商支持 - 支持OpenAI、DeepSeek、MoonShot、通义千问等多家厂商
- 🔄 自动厂商检测 - 根据模型名称自动识别并配置正确的厂商API
- ⚡ 异步并行处理 - 使用异步并发技术提高性能,加快分析速度
📋 目录
📖 项目文档
📚 相关文档
🚀 安装
克隆仓库并安装依赖:
git clone https://github.com/Vistaminc/AuditLuma.git
cd AuditLuma
pip install -r requirements.txt
可选依赖
FAISS向量检索库
默认情况下,AuditLuma会使用一个简单的内置向量存储实现。如果需要处理大型代码库,建议安装FAISS以提高性能:
# CPU版本
pip install faiss-cpu
# GPU版本(支持CUDA)
pip install faiss-gpu
安装FAISS后,系统将自动检测并使用它进行向量存储和检索,显著提高分析大型项目时的性能。
🛠 使用
基本用法:
python main.py -d ./goalfile -o ./reports
命令行参数
| 参数 | 说明 | 默认值 |
|---|---|---|
-d, --directory |
目标项目目录 | ./goalfile |
-o, --output |
报告输出目录 | ./reports |
-w, --workers |
并行工作线程数 | 配置中的max_batch_size |
-f, --format |
报告格式(html/pdf/json) | 配置中的report_format |
--no-mcp |
禁用多智能体协作协议 | 默认启用 |
--no-self-rag |
禁用Self-RAG检索 | 默认启用 |
--no-deps |
跳过依赖分析 | 默认不跳过 |
--no-remediation |
跳过生成修复建议 | 默认不跳过 |
--verbose |
启用详细日志记录 | 默认禁用 |
⚙️ 配置
通过编辑config/config.yaml文件配置系统。主要配置项包括:
模型规范格式
AuditLuma支持使用统一的模型规范格式 model@provider 来指定模型和提供商:
deepseek-chat@deepseek # 指定使用DeepSeek提供商的deepseek-chat模型 gpt-4-turbo@openai # 指定使用OpenAI提供商的gpt-4-turbo模型 qwen-turbo@qwen # 指定使用通义千问提供商的qwen-turbo模型
如果不指定提供商(不使用@符号),系统将自动根据模型名称推断提供商。
配置文件示例
# 默认模型配置
default_models:
code_analysis: "deepseek-chat@deepseek" # 代码分析模型
security_audit: "gpt-4-turbo@openai" # 安全审计模型
remediation: "deepseek-chat@deepseek" # 修复建议模型
summarization: "qwen-turbo@qwen" # 报告总结模型
# Self-RAG配置
self_rag:
enabled: true
embedding_model: "text-embedding-3-small@openai" # 嵌入模型
# 多智能体配置
mcp:
enabled: true
agents:
- name: "orchestrator"
type: "coordinator"
model: "deepseek-chat@deepseek" # 编排智能体使用的模型
- name: "security_analyst"
type: "analyst"
model: "gpt-4-turbo@openai" # 安全分析智能体使用的模型
多厂商支持
AuditLuma支持多家LLM厂商,并能根据模型名称自动检测厂商:
| 模型前缀 | 厂商 |
|---|---|
gpt- |
OpenAI |
deepseek- |
DeepSeek |
qwen- |
通义千问 |
glm-或chatglm |
智谱AI |
baichuan |
百川 |
ollama- |
ollama |
-注意:openai厂商可以对接所有openai格式的中转平台
💻 支持语言
AuditLuma支持分析以下编程语言:
主要语言(包括排名前10)
- Python (.py)
- JavaScript (.js, .jsx)
- TypeScript (.ts, .tsx)
- Java (.java)
- C# (.cs)
- C++ (.cpp, .cc, .hpp)
- C (.c, .h)
- Go (.go)
- Ruby (.rb)
- PHP (.php)
- Lua (.lua)
其他支持的语言
- Rust (.rs)
- Swift (.swift)
- Kotlin (.kt)
- Scala (.scala)
- Dart (.dart)
- Bash (.sh, .bash)
- PowerShell (.ps1, .psm1)
标记和配置语言
- HTML (.html, .htm)
- CSS (.css)
- JSON (.json)
- XML (.xml)
- YAML (.yml, .yaml)
- SQL (.sql)
🏛 架构
AuditLuma使用多代理架构,包含以下组件:
- Agent Orchestrator - 协调工作流中的所有代理
- 代码分析代理 - 分析代码结构并提取依赖关系
- 安全分析代理 - 识别安全漏洞
- 修复建议代理 - 生成针对性漏洞修复方案
- 可视化组件 - 生成直观的报告和依赖关系图
📊 报告格式
AuditLuma支持以下报告格式:
- 📋 HTML报告 - 包含漏洞详情、统计图表和交互式可视化
- 📄 PDF报告 - 适合打印和分享的格式
- 🔄 JSON报告 - 适合进一步处理和集成的机器可读格式
💬 贡献
欢迎贡献代码和建议!请遵循以下步骤:
- Fork 仓库
- 创建特性分支 (
git checkout -b feature/amazing-feature) - 提交更改 (
git commit -m 'Add some amazing feature') - 推送到分支 (
git push origin feature/amazing-feature) - 创建Pull Request
📞 交流方式
-QQ:1047736593
支持与赞赏
如果您觉得AuditLuma对您有帮助,欢迎通过以下方式支持我们:
- 您的赞助将用于帮助我们持续改进和完善 AuditLuma!
📜 许可证
MIT
Dev Tools Supporting MCP
The following are the main code editors that support the Model Context Protocol. Click the link to visit the official website for more information.










