Refactor Knowledge Base Schema to Fix Type Regression

The AI query tool (npm run ai:query) fails for --type example because the knowledge base creation script does not correctly assign a high-level type to items originating from the /examples directory. The current implementation uses type to store the specific code construct (class, method), creating a conflict.

This ticket refactors the knowledge base schema to solve this regression and make the data model more robust.

Tasks:

1. Modify buildScripts/ai/createKnowledgeBase.mjs:

   • For each chunk generated from the JSDoc output (all.json), add a type property (src, example, etc.) based on its source file path.
   • Rename the existing type property (which currently holds class, method, etc.) to kind to avoid conflicts.
   • For non-JSDoc chunks (guides, releases, tickets), ensure they also have both a type and a kind property (where both can have the same value, e.g., type: 'guide', kind: 'guide').

2. Verify buildScripts/ai/queryKnowledgeBase.mjs:

   • Confirm that the existing query logic, which filters on the type property, works correctly with the new schema without modification.