Add: SRAO Domain Modeler — context engineering for multi-agent industrial systems

[beixuan577]

Context Engineering for Multi-Agent Systems

While most context engineering focuses on LLM prompts, there's a higher-order context problem: domain context for agent systems.

The SRAO Domain Modeler approaches this systematically:

Domain Context Dimensions

1. Entity context: Complete concept dictionaries with types, attributes, and relationships (10+ entities per industry)
2. Workflow context: Standardized workflow templates with serial/parallel/branch semantics
3. Cross-industry context: Reuse matrix identifying shared concepts across 5 industries
4. Stakeholder context: SRM format capturing constraints, resources, and acceptance criteria

Concrete example (Manufacturing)

Concept Dictionary:
  Order → WorkOrder → ProductionLine → Equipment → Sensor → Alert → MaintenanceTicket

Workflow Template:
  ParseOrder → [CheckInventory || CalculateCapacity] → OptimizeSchedule → GenerateWorkOrder
  
Reuse Matrix:
  "Equipment" concept → maps to "WindTurbine" (energy), "MedicalDevice" (healthcare), "FarmMachine" (agriculture)

Why this matters for context engineering

When building multi-agent systems, the quality of domain context directly determines agent effectiveness. SRAO provides structured, reusable context templates rather than ad-hoc prompt engineering.

Full framework: https://github.com/beixuan577/SRAO-Framework