This is an automated donwport of ABAP MCP SDK. This is a best effort donwport using the awesome downport features of abaplint. Unfortunately I have no system with a release below 7.50 available, therefore bugs and syntax issues are expected. Downport target is 7.02 which I assume should work.
If you face issues feel free to open an issue.
Discord - for all ABAP-AI repos.
ABAP implementation of the Model Context Protocol based on version 2025-03-26.
Note that this is only a server implementation as I currently see no realistic way to support SSE streaming which is mandatory for the client.
See Overview in docs folder.
Currently implemented:
- Prompts
- Resources
- Tools
No SSE support --> no notifications and therefore no support of listChanged, subscribe and logging.\
Not implmented:
- Sampling - considered for the future, open an issue if you have a use case that would benefit from it
- Completions - questionable from a performance point of view with ABAP
- Roots - likely irrelevant for ABAP
- Further notifications not relevant without SSE: Cancellation, Ping, Progress, Logging
Non-compliance to the spec:
- Authorization standard cannot be implemented in ABAP --> we rely on ABAP capabilties instead
See currently open issues for details. In general the following topics are prioritized:
- Full test suite especially covering session handling and error scenarios
- Demonstrate ways to use OAuth reducing challenges with only standard compliant clients
- Cleanup batch job / report to delete outdated sessions
- Improve logging by increasing the logged scope
- Add options to record requests & responses for debugging purposes
- Alternatve composable server - instead of writing a full server just configure a combination of standalone implementations for better reuse
- Examples, documentation, etc.
Feel free to:
- Open issues for bug reports
- Use Discord for discussions
- Suggest improvements
- Pull requests are welcome but ensure they are tested and follow the overall style, if in doubt open an issue first
Part of the coding was generated using AI, mostly with Sonnet 3.7 and GitHub Copilot. Detailed documentation is largely AI generated - only high-level reviewed.
