Evaluate jeanibarz/knowledge-base-mcp-server for the Comparisons landscape

[jmchilton]

Context

The reader-facing Related Projects page was removed because it mixed pattern-level prior art with projects inherited from the Statistical Genomics Foundry's neighborhood. We should not move that inventory wholesale into Comparisons.

One project merits a focused follow-up: jeanibarz/knowledge-base-mcp-server. It is a runtime knowledge-base system exposed through MCP and a CLI, but its current feature set includes source citations, provenance-tagged imports, evidence packets, retrieval evaluations, and eval gates. Those features may challenge overly broad claims that runtime knowledge-base systems lack traceability or enforcement.

Question

Does this project occupy or narrow any part of the Foundry Pattern's claimed niche strongly enough to deserve a named entry in Comparisons, or is it adequately covered by the existing MCP/RAG categories?

Evaluate

• Its source model: what is authoritative, human-readable, and editable?
• When grounding happens: authoring time, build time, or runtime.
• Whether it produces a portable agent artifact or only serves a live knowledge base.
• What its citations and provenance-tagged imports actually guarantee.
• Whether its evidence packets and eval gates enforce correctness, retrieval quality, or something narrower.
• How portable its knowledge and behavior are across runtimes.
• Whether a named entry would add a genuinely new comparison rather than another implementation inventory item.

Done when

• Claims are checked against dated primary sources.
• We record one explicit outcome:
  • add a named Comparisons entry;
  • strengthen an existing MCP/RAG entry without naming the project; or
  • make no documentation change.
• Any proposed comparison states the strongest shared ground first and names the remaining distinction precisely.
• We do not recreate a generic “knowledgebase-MCP” cluster or reintroduce the removed Related Projects inventory.