Epic: Blackboard Architecture Evolution — Research-Identified Future Improvements

[mdproctor]

Overview

Captures future improvements to the blackboard architecture identified through academic literature review (2024–2025 papers on LLM-based blackboard multi-agent systems, BB1, DARBS, reactive scheduling research).

These are not for the current port sprint. The port (casehub-blackboard/ module, issue #76) comes first. This epic tracks what to evolve toward once the foundation is solid.

Research basis

• Exploring Advanced LLM Multi-Agent Systems Based on Blackboard Architecture (arXiv 2507.01701)
• LLM-Based Multi-Agent Blackboard System for Information (arXiv 2510.01285)
• BB1 (Hayes-Roth 1985) — meta-level control knowledge sources
• DARBS — distributed blackboard, client/server knowledge source model
• Reactive scheduling with reason maintenance (ScienceDirect)

Child issues

Issue	Area
#78	Meta-level control knowledge sources (BB1-style stall detection + strategy switching)
#79	Dual-space blackboard — private agent scratchpad separate from public shared workspace
#80	Memory stratification — working, episodic, and semantic panels on CaseFile
#81	Hierarchical blackboard panels — named regions / zones of the shared workspace
#82	Dynamic agent selection — capability and confidence-based activation beyond key existence
#83	Reason maintenance — structured justification tracking for TaskDefinition activations

Relationship to current work

The casehub-blackboard/ port (issue #76, epic #30) is the prerequisite. These improvements extend and deepen the architecture once that foundation exists.

Note: Memory stratification (#80) and hierarchical panels (#81) are closely related — evaluate together before implementing either.