feat(ai): Nativize memory core ingestion and standardize AST graph logic

Architectural Change: ChromaDB Deprecation & SQLite Native Migration

Successfully completed the bridge moving all temporal Memory Core execution out of legacy Python ChromaDB infrastructure and natively into neo-sqlite:

1. Rescued ~8,000 legacy episodic memories (730 sessions) via extracting them from Chroma, manually aligning them into our strict 4096D format via MLX Qwen3 embeddings, and importing them directly into the internal SQLite matrix neo_agent_sessions_data.
2. Created high-performance SQLite importer/rescue utilities in /buildScripts/ai/ to bypass MCP boundaries during large volumetric data ingestions.
3. Activated REM-Sleep synthesis natively pointing solely to local SQLite Edge logic via DreamService.

AST Node Taxonomy Consistency

1. Stripped @class and @extends annotations universally across all Node execution components within buildScripts/ai/*.
2. Replaced documentation headers universally with @module following the Anchor & Echo knowledgebase enhancement strategy. This mitigates AST topology contamination, ensuring the FileSystemIngestor natively parses scripts into isolated File structures containing isolated Function nodes rather than grouping them hierarchically as pseudo-classes inside the native knowledge graph.