AI interface for DOORS Classic

As more and more AI is pushing our life of work, data exchange between DOORS Classic and AI Systems are a must to provide the benefits from huge Requirements Management Database and their knowledge-base!
As I'm asked internally from our central department multiple times for an interface connecting DOORS Classic and our internal AI system(s), I would like to ask for a possible solution.
Best will be to use standardized protocols/implementations like MCP or similar (see references) which helps to get an (readonly?) interface for the DOORS (Classic) Database.
 

Thinking about the possible existing interfaces for DOORS, i.e. Client-batchmode, OSLC, OSLC DXL-Services, I think the implementation must use an environment with the least possible complexity. I.e. OSLC but not OSLX DXL-Services (as these are single-tasking only!).

An example Use Case can be:
A user is editing some source-code (i.e. in VSCode) and would like to know some related information of the connected Requirement(s). Using a command interface and asking the AI (i.e. Copilot) and want to get the related Requirements (Object-Id's, Headings and maybe Object Text) or a summary of the related requirements.

Another example may be:
A user is implementing some settings-dialog for a software application and would like to know all setting-related requirements of the project. Then a list with all related (also baselined related information) is presented and the the user would like to get a summarized list of short sentences which can be taken for the dialog-box to show the setting-name to quickly implement the dialog-box.

Also (maybe) the user would like to "Show me all references in the Projects A, B and C of the maximum speed of the train Superfast driving through an high-speed switch." to get an overview how to dimension some setup.

References:
* https://en.wikipedia.org/wiki/Model_Context_Protocol
* https://modelcontextprotocol.io/docs/getting-started/intro