Architect AI Knowledge Base as MCP Server

This epic addresses the need to migrate our script-based AI knowledge base tools (buildScripts/ai/*.mjs) into a robust, professional-grade Model Context Protocol (MCP) server.

The primary goal is to create a persistent, stateful server that provides a formal API for all knowledge base interactions (building, embedding, and querying). This moves us away from ad-hoc scripts and toward the professionalized AI tooling outlined in the project ROADMAP.md.

Priority & Methodology

This is a critical-path initiative for the core team (Tobi & Gemini duo). It is not a candidate for Hacktoberfest or external contribution.

We will employ a rapid and agile development approach. The scope and API specification are expected to evolve as we build and test the server. This flexibility is key to quickly delivering a solution that effectively unblocks our ability to provide AI agents with deep project context.

Top-Level Items

Phase 1: Scaffolding & Core API Definition

• Goal: Establish the server's foundation and define the initial API for knowledge base interactions.
• Sub-Tasks:
  • ticket-kb-design-api.md: Design the comprehensive API for the Knowledge Base MCP Server.
  • ticket-kb-scaffold-server.md: Scaffold the server structure based on the mcp-stdio.mjs pattern.

Phase 2: API Implementation

• Goal: Build out the backend logic to connect the OpenAPI endpoints to live service functions.
• Sub-Tasks:
  • ticket-kb-implement-tool-service.md: Implement the core tool service to dynamically load tools from the OpenAPI spec.
  • ticket-kb-implement-healthcheck-service.md: Implement the healthcheck service to verify the connection to ChromaDB.
  • ticket-kb-implement-delete-db-service.md: Implement the service to delete the ChromaDB collection.
  • ticket-kb-implement-query-service.md: Implement the service to query documents from the knowledge base.
  • ticket-kb-implement-sync-service.md: Implement the service to build and embed the knowledge base content.
  • ticket-kb-implement-document-services.md: Implement services to list and retrieve individual documents.

Phase 3: Refactoring & Cleanup

• Goal: Improve the modularity and maintainability of the server code.
• Sub-Tasks:
  • ticket-kb-refactor-config.md: Decouple the server from the shared buildScripts configuration.
  • ticket-kb-separate-create-embed.md: Separate the monolithic sync service into distinct create and embed tools.
  • ticket-kb-review-and-correct-database-service.md: Review and correct the databaseService implementation to ensure full feature parity.
  • ticket-kb-enhance-openapi-examples.md: Enhance the OpenAPI spec with tool usage examples from AGENTS.md.
  • ticket-kb-enhance-tool-manuals.md: Expand tool descriptions into comprehensive in-spec manuals.
  • ticket-kb-refactor-tool-service.md: Refactor toolService.mjs to reduce code duplication.
  • ticket-kb-resilient-tool-service.md: Make toolService resilient to server-prefixed tool names.
  • ticket-kb-openapi-driven-args.md: Dynamically determine argument passing strategy from OpenAPI spec.

Phase 4: Bugfixes

• Goal: Address bugs discovered during testing.
• Sub-Tasks:
  • bug-kb-healthcheck-unstructured-content.md: Fix the healthcheck tool to return structured content matching its schema.
  • bug-kb-query-documents-input-mismatch.md: Fix the input parameter mismatch for the query_documents tool.