• 中文文档
  • 功能特性
  • 系统要求
  • 快速开始
  • 项目结构
  • 检查模块详解
  • 评分体系
  • 报告生成
  • 修复功能
  • CI/CD 集成
  • AI 深度审计
  • 参与贡献
• English Documentation
  • Features
  • Requirements
  • Quick Start
  • Project Structure
  • Check Modules
  • Scoring System
  • Report Generation
  • Fix Functionality
  • CI/CD Integration
  • AI-Assisted Audit
  • Contributing

• Python 3.8+
• Windows 7+ / Linux / macOS
• 管理员权限（用于部分系统级检测）
• 依赖库：
  • customtkinter >= 5.2.0
  • Pillow >= 10.0.0
  • pyyaml >= 6.0.1
  • psutil >= 5.9.0

1. 从 build/windows/dist/ 目录下载 焦糖布丁v4.0.exe
2. 双击运行可执行文件
3. 在界面中选择 OpenClaw 安装目录
4. 勾选需要检测的安全维度
5. 点击"开始全面检测"按钮启动安全扫描
6. 查看检测结果并导出报告

1. 克隆项目到本地
2. 安装依赖：

   pip install -r requirements.txt

3. 运行主程序：

   python src/main.py

• 安全审计
• 自动加固
• Skills技能安全扫描
• 完整性检查
• 隐私检查
• 行为检测检查

以集成本地ollama模型和云端模型的深度审计功能

各模块按权重加权计算总评分，输出风险等级：

扫描完成后，您可以导出两种格式的报告：

HTML 报告采用插画风格设计，包含完整的检测结果和修复建议，方便您快速了解系统安全状态。

对于龙虾安全守卫功能，您可以单独导出HTML报告，包括：

• 安全审计报告
• 自动加固报告
• 技能扫描报告
• 文件完整性检查报告
• 隐私检查报告
• 行为检测检查报告

对于检测到的安全问题，您可以：

1. 在结果表格中选择需要修复的风险项（支持多选）
2. 点击"修复选中风险"按钮
3. 在弹出的对话框中查看修复操作详情
4. 点击"执行修复"按钮进行修复
5. 修复完成后，工具会自动重新扫描以更新结果

修复操作说明：

• 对于低风险和中风险的问题，工具会尝试自动修复
• 对于高风险和严重风险的问题，工具会提供修复建议，需要人工确认后执行

焦糖布丁提供 AI 深度审计功能，利用 AI 对检测结果进行更深入的分析：

1. 点击界面中的"AI 深度审计"按钮
2. 工具会将检测结果生成AI提示词模板，然后将模板发给AI来进行详细的分析
3. AI 会提供更详细的安全建议和修复方案
4. 您可以根据 AI 的建议进行进一步的安全加固

工具修复方案和内容仅供参考，由于传播、利用此工具所提供的信息而造成的任何直接或者间接的后果及损失，均由使用者本人负责，作者不为此承担任何责任。

MIT License)

• Python 3.8+
• Windows 7+ / Linux / macOS
• Administrator privileges (for system-level checks)
• Dependencies:
  • customtkinter >= 5.2.0
  • Pillow >= 10.0.0
  • pyyaml >= 6.0.1
  • psutil >= 5.9.0

1. Download 焦糖布丁v4.0.exe from the build/windows/dist/ directory
2. Double-click to run the executable
3. Select the OpenClaw installation directory in the interface
4. Check the security dimensions to scan
5. Click "开始全面检测" (Start Full Scan) to launch the security scan
6. View the results and export reports

1. Clone the project to your local machine
2. Install dependencies:

   pip install -r requirements.txt

3. Run the main program:

   python src/main.py

• Security audit
• Auto hardening
• Skills security scanning
• Integrity checking
• Privacy checking
• Behavior detection checking

Deep audit functionality with local ollama model and cloud model integration

After scanning, you can export reports in two formats:

The HTML report uses an illustration style design, including complete scan results and fix suggestions, making it easy to quickly understand the system security status.

For Lobster Security Guard features, you can export separate HTML reports including:

• Security audit report
• Auto hardening report
• Skill scanning report
• File integrity check report
• Privacy check report
• Behavior detection check report

For detected security issues, you can:

1. Select the risk items to fix in the results table (multi-select supported)
2. Click the "修复选中风险" (Fix Selected Risks) button
3. View the fix operation details in the pop-up dialog
4. Click "执行修复" (Execute Fix) to perform the fix
5. After fixing, the tool will automatically rescan to update the results

Fix Operation Notes:

• For low and medium risk issues, the tool will attempt to fix them automatically
• For high and critical risk issues, the tool will provide fix suggestions that require manual confirmation

焦糖布丁 provides AI-assisted audit functionality to perform deeper analysis of scan results:

1. Click the "AI 深度审计" (AI Deep Audit) button in the interface
2. The tool will generate an AI prompt template from the scan results and send it to AI for detailed analysis
3. AI will provide more detailed security suggestions and fix solutions
4. You can follow AI's suggestions for further security hardening

The tool's fix solutions and content are for reference only. The author is not responsible for any direct or indirect consequences and losses caused by the dissemination or use of the information provided by this tool. The user assumes full responsibility.

MIT License