ABAP Transport Analyzer MCP Server
Enables AI-driven analysis of SAP transport requests, including metadata retrieval, code diff analysis, and risk detection for ABAP objects.
README
AI-Powered ABAP Transport Analyzer - MCP Server
An intelligent Model Context Protocol (MCP) server that automates SAP transport code review, change analysis, and risk detection using the SAP ADT (ABAP Development Tools) REST API.
Features
- Transport Metadata Retrieval - Fetch transport request details, owner, status, and object list
- Automated Code Diff Analysis - Generate unified diffs for ABAP objects (Classes, Reports, Interfaces, Function Modules)
- Risk Detection - Identify security risks, breaking changes, and code quality issues
- Natural Language Summaries - LLM-optimized structured analysis reports
- Version Management - Compare active code against previous versions from SAP
- Error Handling - Graceful handling of authorization, network, and parsing errors
Prerequisites
- Node.js 18+
- SAP S/4HANA system with ADT enablement (SAP Note 2162659 or later)
- SAP User Account with authorizations:
S_TRANSPRT(Transport Management)S_DEVELOP(ABAP Development)
- Environment Variables for SAP connection (see Configuration section)
Installation
1. Clone or Download Project
git clone <repository-url>
cd my-abap-mcp-server
2. Install Dependencies
npm install
3. Configure Environment
Copy .env.example to .env and update with your SAP system details:
cp .env.example .env
Edit .env:
SAP_HOST=https://your-sap-system.example.com:44300
SAP_CLIENT=100
SAP_USER=your_abap_user
SAP_PASSWORD=your_secure_password
4. Build TypeScript
npm run build
5. Start the Server
# Development mode (with auto-reload)
npm run dev
# Production mode
npm start
Configuration
Environment Variables
| Variable | Description | Example | Required |
|---|---|---|---|
SAP_HOST |
Full URL to SAP S/4HANA system with ADT enabled | https://saphana.corp.com:44300 |
✅ Yes |
SAP_CLIENT |
SAP client number | 100 |
Optional (default: 100) |
SAP_USER |
ABAP user with S_TRANSPRT and S_DEVELOP auth | DEVELOPER01 |
✅ Yes |
SAP_PASSWORD |
User password (store securely in vault for production) | P@ssw0rd123 |
✅ Yes |
SSL Certificate Handling
If your SAP system uses self-signed certificates, the server currently accepts them. For production, replace the HTTPS agent configuration in src/index.ts:
// Current (development only)
httpsAgent: new https.Agent({ rejectUnauthorized: false })
// For production with trusted CA
httpsAgent: new https.Agent({
ca: fs.readFileSync('/path/to/ca-bundle.pem')
})
Available Tools
1. get_transport_metadata
Retrieves transport request header information and object list.
Input:
{
"transportId": "S4HK900123"
}
Output:
{
"transportId": "S4HK900123",
"description": "Fix pricing calculation in SD module",
"owner": "ABAPDEV",
"status": "Released",
"createdDate": "2026-06-28T10:30:00Z",
"targetSystem": "PROD",
"objectCount": 3
}
2. analyze_transport_changes
Performs detailed code diff analysis and risk assessment for all objects in a transport.
Input:
{
"transportId": "S4HK900123"
}
Output: Markdown-formatted report including:
- Executive summary (objects analyzed, risk counts)
- Per-object change analysis with diffs
- Risk factor classification (HIGH/MEDIUM/LOW)
- Recommendations and mitigation steps
Usage Examples
Integrate with Claude or other LLM
# Example with Claude API
import anthropic
client = anthropic.Anthropic()
response = client.messages.create(
model="claude-3-5-sonnet-20241022",
max_tokens=2048,
tools=[
{
"name": "get_transport_metadata",
"description": "Get transport request metadata",
"input_schema": { ... }
},
{
"name": "analyze_transport_changes",
"description": "Analyze code changes in transport",
"input_schema": { ... }
}
],
messages=[
{
"role": "user",
"content": "What changed in transport S4HK900123?"
}
]
)
Direct CLI Usage (with jq)
# Test connection
curl -X POST http://localhost:3000/tools/get_transport_metadata \
-H "Content-Type: application/json" \
-d '{"transportId": "S4HK900123"}'
Supported ABAP Object Types
The analyzer extracts and analyzes these ABAP object types:
CLAS- ABAP ClassesPROG- ABAP Reports/ProgramsINTF- ABAP InterfacesFUGR- Function GroupsFUNC- Function ModulesTABL- Database Tables (schema changes)VIEW- Database ViewsTYPE- Type DefinitionsENPD- Enhancement PointsENHS- Enhancements
Risk Detection Rules
The analyzer identifies the following risk categories:
HIGH Severity
- Missing AUTHORITY-CHECK for database modification operations (INSERT/UPDATE/DELETE)
- API breaking changes (visibility changed to PRIVATE)
- Modifications to standard SAP objects
MEDIUM Severity
- Hardcoded numeric values or hex constants
- Significant code deletions (>20 lines, >70% of changes)
- New external method/function calls with potential dependency issues
LOW Severity
- Code style improvements
- Comment updates
API Error Responses
| Error | Status | Cause | Solution |
|---|---|---|---|
| Authorization Failed | 401 | Invalid credentials or insufficient SAP authorization | Verify SAP user has S_TRANSPRT, S_DEVELOP authorities |
| Transport Not Found | 404 | Transport ID doesn't exist or is not accessible | Confirm transport ID is correct and released |
| XML Parse Error | 500 | ADT response format unexpected | Check SAP system release compatibility |
| Timeout | 504 | SAP backend unresponsive (>15s) | Check SAP system health, retry request |
Testing
Run the test suite:
# Run all tests
npm test
# Run with coverage
npm test -- --coverage
# Run specific test file
npm test -- tests/mcp-server.test.ts
# Watch mode (auto-rerun on file changes)
npm test -- --watch
Logging
The server writes debug logs to debug.log in the project root. This file is NOT part of the MCP protocol output to prevent corruption.
Monitor logs in real-time:
# On Windows PowerShell
Get-Content debug.log -Wait -Tail 20
# On macOS/Linux
tail -f debug.log
Logs include:
- Transport metadata parsing details
- XML structure analysis
- Object extraction trace
- HTTP request/response summaries
Deployment
Docker
Build and run in a Docker container:
docker build -t abap-mcp-server .
docker run -e SAP_HOST=https://... -e SAP_USER=... -e SAP_PASSWORD=... abap-mcp-server
Kubernetes
Deploy to K8s cluster (see k8s-deployment.yaml):
kubectl apply -f k8s-deployment.yaml
Cloud Platforms
- AWS Lambda - Package with Layers for node_modules
- Azure Functions - Use Node.js runtime
- Google Cloud Run - Containerize and deploy
Development
Project Structure
my-abap-mcp-server/
├── src/
│ └── index.ts # Main MCP server implementation
├── tests/
│ ├── mcp-server.test.ts # TypeScript unit tests
│ └── mcp-server.test.js # JavaScript unit tests
├── dist/ # Compiled JavaScript (generated)
├── .env.example # Environment template
├── package.json # Dependencies & scripts
├── tsconfig.json # TypeScript configuration
├── jest.config.js # Test runner configuration
├── nodemon.json # Auto-reload configuration
└── README.md # This file
Build Commands
# Compile TypeScript to JavaScript
npm run build
# Start development server with auto-reload
npm run dev
# Run production server
npm start
# Run tests
npm test
Code Style
- Language: TypeScript with strict mode enabled
- Formatting: Follows Node.js conventions
- Linting: (TODO: Add ESLint)
Troubleshooting
Connection Issues
Problem: "Missing SAP connection credentials in .env file"
Solution:
- Verify
.envfile exists in project root - Check all required variables are set:
SAP_HOST,SAP_USER,SAP_PASSWORD - Ensure no trailing spaces or quotes in
.envvalues
Problem: "Authorization Failed: Invalid SAP credentials"
Solution:
- Verify SAP user password is correct
- Check user has S_TRANSPRT and S_DEVELOP authorizations
- In SAP, go to SUIM transaction and verify role assignments
Problem: "Transport Not Found: Transport ID does not exist"
Solution:
- Confirm transport ID is spelled correctly (case-sensitive in some systems)
- Verify transport is released (check status in SE10/SE09)
- Ensure user has authorization to view the transport
Performance Issues
Problem: Analysis takes >5 seconds per object
Solutions:
- Check SAP system performance (SE30, SM50)
- Verify network latency to SAP system (ping test)
- Consider implementing caching for frequently accessed transports
- Analyze smaller transports first (split large ones)
XML Parsing Errors
Problem: "Failed to parse XML response for transport"
Solution:
- Check
debug.logfor the actual XML structure returned - Verify SAP system release (S/4HANA 2020 or later recommended)
- Review SAP Note 2162659 for ADT configuration
Security Considerations
⚠️ Important Security Notes:
-
Credential Management
- Never commit
.envfile to version control - Use environment variables or secret vaults in production
- Rotate SAP credentials regularly
- Use OAuth2 if available (future enhancement)
- Never commit
-
Data Privacy
- ABAP source code retrieved may contain sensitive business logic
- Ensure logs are protected with appropriate access controls
- Only share diffs with authorized personnel
- Consider data classification policies before transmission
-
Network Security
- Always use HTTPS for SAP connections
- Validate SSL certificates in production (not disabled)
- Firewall restrict access to MCP server endpoints
- Implement rate limiting for production use
-
Authorization
- Audit user access to transports regularly
- Limit MCP server access to authorized LLM agents
- Monitor for suspicious transport analysis patterns
- Log all tool invocations for compliance
Roadmap
- [ ] Support for multiple SAP systems (multi-tenant)
- [ ] OAuth2 authentication support
- [ ] Transport comparison (side-by-side analysis)
- [ ] Caching layer for performance optimization
- [ ] Advanced risk rules (custom, configurable)
- [ ] Batch transport analysis
- [ ] HTML/PDF report generation
- [ ] Slack/Teams integration for alerts
- [ ] Kubernetes Helm charts
- [ ] Performance metrics & monitoring (Prometheus)
Support & Contribution
- Issues: Report bugs via GitHub Issues
- Questions: Open Discussions tab
- Contributions: See CONTRIBUTING.md for guidelines
- License: ISC
References
Last Updated: 2026-07-05
Version: 1.0.0
Maintainer: ABAP Development Team
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。