Unclassified
Core Spine Agents
The Foundation of the Agent Network
Core Spine Agents are the five system-level agents that form the foundation of IntegrateWise. They work together to coordinate all interactions, manage data, discover capabilities, and route work efficiently.
Key Insight: Without these core agents, the Spine Agent Network cannot function. They are always active and provide the essential infrastructure for all other agents.
The Five Core Agents
1. Twin Agent
Your Digital Representative
The Twin is your permanent digital identity within IntegrateWise. It maintains your context, preferences, knowledge, and work patterns across all interactions.
Key Responsibilities
- Identity Management - Maintains your digital identity across the system
- Context Coordination - Assembles and manages context for all your work
- Agent Orchestration - Coordinates other agents to complete your requests
- Continuity - Ensures seamless experience across all interactions
- Personalization - Learns and adapts to your unique needs and patterns
How Twin Works
User Request → Twin Analysis → Agent Selection → Context Assembly → Execution → Result Delivery
Example: When you ask "What's the status of the Acme deal?", Twin:
- Understands you're asking about a sales opportunity
- Identifies the need for CRM data and previous interactions
- Selects appropriate agents (Discovery Agent, CRM connectors)
- Assembles context about Acme from Spine Fabric
- Coordinates agents to gather and synthesize information
- Delivers comprehensive answer with full context
User Experience
- Single Interface - Talk to Twin, not to individual agents
- Consistent Context - Twin remembers your preferences and work history
- Intelligent Delegation - Twin automatically routes work to the right agents
- Proactive Assistance - Twin suggests relevant actions based on context
2. Schema Agent
The Data Architect
The Schema Agent is the guardian of IntegrateWise's canonical data model, ensuring all entities, events, and contexts conform to defined schemas.
Key Responsibilities
- Schema Validation - Validates all incoming data against canonical schemas
- Schema Synthesis - Automatically generates and evolves schemas from data
- Entity Mapping - Maps between different data models (external → Spine canonical)
- Event Extraction - Extracts structured events from unstructured data
Supported Entity Types
| Category | Examples |
|---|---|
| Business Entities | Customer, Account, Opportunity, Contact, Lead |
| Work Entities | Task, Project, Deal, Ticket, Meeting |
| Communication | Email, Message, Call, Chat |
| Content | Document, Note, File, Knowledge Base |
| Financial | Invoice, Payment, Order, Contract |
| Events | Created, Updated, Deleted, Completed |
Schema Evolution
The Schema Agent supports:
- Automatic Schema Inference - Derives schemas from data patterns
- Schema Versioning - Manages schema evolution over time
- Conflict Resolution - Handles schema conflicts and merges
- Custom Schemas - Supports tenant-specific schema extensions
3. Coworker Agent
Your Collaboration Partner
The Coworker Agent works alongside humans on shared tasks, providing intelligent assistance for collaborative work.
Key Responsibilities
- Task Collaboration - Works with humans on shared tasks and projects
- Context Sharing - Maintains shared context for team work
- Handoff Management - Smoothly transitions work between team members
- Coordination - Helps coordinate multi-person workflows
Collaboration Patterns
-
Human + Coworker
Human starts task → Coworker assists → Human reviews → Coworker refines → Task complete -
Multiple Humans + Coworker
Team discussion → Coworker captures context → Individual work → Coworker coordinates → Team sync -
Coworker + Specialized Agents
Complex task → Coworker decomposes → Subtasks to agents → Coworker aggregates → Final delivery
Use Cases
- Document Creation - Collaborate on documents with AI assistance
- Project Planning - Plan projects with intelligent suggestions
- Code Review - Review code with AI-powered analysis
- Meeting Preparation - Prepare for meetings with comprehensive briefings
4. Discovery Agent
The Information Finder
The Discovery Agent is responsible for finding relevant context, tools, capabilities, and information to support agent operations.
Key Responsibilities
- Capability Discovery - Finds agents and tools that can perform needed tasks
- Context Discovery - Identifies relevant information and context for requests
- Tool Discovery - Discovers available tools and their capabilities
- Route Optimization - Determines the most efficient path for work execution
Discovery Mechanisms
-
Capability-Based Discovery
Request: "Analyze this market" Discovery: Finds agents with "market_analysis" capability Result: Returns Market Research Agent -
Context-Based Discovery
Context: Working on Salesforce opportunity Discovery: Finds CRM tools and Sales agents Result: Returns Salesforce connector + Lead Agent -
Pattern-Based Discovery
Pattern: User frequently asks for customer health scores Discovery: Identifies Customer Success agents and metrics Result: Returns Account Health Agent + relevant dashboards
Discovery Sources
- Agent Registry - Database of all available agents and their capabilities
- Tool Registry - Database of all connected tools and their functions
- Context Index - Index of all context and relationships in Spine Fabric
- Usage Patterns - Historical patterns of user behavior and preferences
5. Triage Agent
The Work Router
The Triage Agent is responsible for routing requests, tasks, and work to the appropriate agents, systems, or humans based on priority, capability, and context.
Key Responsibilities
- Request Prioritization - Determines urgency and importance of incoming requests
- Agent Selection - Routes work to the most appropriate agent
- Load Balancing - Distributes work across available agents
- Exception Handling - Manages edge cases and error conditions
- Queue Management - Maintains work queues and processing order
Triage Decision Factors
| Factor | Description | Example |
|---|---|---|
| Urgency | Time sensitivity of the request | "ASAP" vs "When convenient" |
| Importance | Business impact of the request | Customer escalation vs routine task |
| Complexity | How complex the task is | Multi-step workflow vs simple query |
| Capabilities | What capabilities are needed | Requires CRM access + analysis |
| Context | Current user context | Working on specific customer or project |
| Availability | Which agents are available | Agent load and readiness |
| Dependencies | What the task depends on | Waiting for other tasks to complete |
Triage Workflow
1. Request Received → Analyze request type and context
2. Priority Assessment → Determine urgency and importance
3. Capability Matching → Find agents with required capabilities
4. Availability Check → Verify agent availability
5. Route Decision → Select optimal agent or workflow
6. Handoff → Transfer work with full context
7. Monitor → Track execution and handle exceptions
Agent Relationships
How Core Agents Work Together
┌─────────────────────────────────────────────────────────────┐
│ CORE AGENT COLLABORATION │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │ TWIN │ │ DISCOVERY │ │ SCHEMA │ │
│ │ (Orchestrator)│ │ (Finder) │ │ (Architect) │ │
│ └──────┬───────┘ └──────┬───────┘ └──────┬───────┘ │
│ │ │ │ │
│ │ ┌──────────────▼──────────────┐ │ │
│ │ │ WORK REQUEST │ │ │
│ │ │ ┌─────────────────────────┐│ │ │
│ │ │ │ What needs to be done? ││ │ │
│ │ │ │ What capabilities needed?││ │ │
│ │ │ └─────────────────────────┘│ │ │
│ │ └──────────────┬──────────────┘ │ │
│ │ │ │ │
│ ▼ ▼ ▼ │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │ COWORKER │ │ TRIAGE │ │ Validates │ │
│ │ (Collaborator)│ │ (Router) │ │ Data & │ │
│ └──────────────┘ └──────────────┘ │ Schemas │ │
│ └──────────────┘ │
│ │
│ ┌─────────────────────────────────────────────────────────┐ │
│ │ EXECUTION FLOW │ │
│ │ │ │
│ │ Twin → Discovery → Triage → Specialized Agent → Execution │ │
│ │ ↓ ↓ ↓ │ │
│ │ Context Capabilities Priority │ │
│ └─────────────────────────────────────────────────────────┘ │
│ │
└─────────────────────────────────────────────────────────────┘
Example: Multi-Agent Collaboration
User Request: "Prepare a comprehensive report on our Q2 sales performance"
- Twin receives request, understands intent
- Discovery finds relevant agents (Data Analysis, Report Generation, Sales Context)
- Schema validates data structures for sales metrics
- Triage prioritizes and routes to Report Generation Agent
- Coworker assists with data gathering and formatting
- Specialized Agents execute their parts (data analysis, chart generation, etc.)
- Twin aggregates results and delivers final report
Technical Details
Agent Architecture
All Core Agents share a common architecture:
┌─────────────────────────────────────────────────────────────┐
│ CORE AGENT STRUCTURE │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌─────────────────────────────────────────────────────────┐ │
│ │ CONTROL LAYER │ │
│ │ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │ │
│ │ │ Intent │ │ Context │ │ Governance │ │ │
│ │ │ Parser │ │ Manager │ │ Engine │ │ │
│ │ └─────────────┘ └─────────────┘ └─────────────┘ │ │
│ └─────────────────────────────────────────────────────────┘ │
│ │
│ ┌─────────────────────────────────────────────────────────┐ │
│ │ EXECUTION LAYER │ │
│ │ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │ │
│ │ │ Task │ │ Handoff │ │ Monitor │ │ │
│ │ │ Executor │ │ Manager │ │ Engine │ │ │
│ │ └─────────────┘ └─────────────┘ └─────────────┘ │ │
│ └─────────────────────────────────────────────────────────┘ │
│ │
│ ┌─────────────────────────────────────────────────────────┐ │
│ │ INTEGRATION LAYER │ │
│ │ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │ │
│ │ │ Spine │ │ MCP │ │ Other │ │ │
│ │ │ Fabric │ │ Pool │ │ Agents │ │ │
│ │ └─────────────┘ └─────────────┘ └─────────────┘ │ │
│ └─────────────────────────────────────────────────────────┘ │
│ │
└─────────────────────────────────────────────────────────────┘
Capability Matrix
| Capability | Twin | Schema | Coworker | Discovery | Triage |
|---|---|---|---|---|---|
| Context Assembly | ✅ | ❌ | ✅ | ✅ | ✅ |
| Task Execution | ✅ | ❌ | ✅ | ❌ | ✅ |
| Data Validation | ❌ | ✅ | ❌ | ❌ | ❌ |
| Capability Discovery | ❌ | ❌ | ❌ | ✅ | ❌ |
| Work Routing | ✅ | ❌ | ❌ | ✅ | ✅ |
| Collaboration | ✅ | ❌ | ✅ | ❌ | ❌ |
| Handoff Management | ✅ | ❌ | ✅ | ❌ | ✅ |
Configuration
Core Agent Settings
All Core Agents are always active and cannot be disabled. However, you can configure their behavior:
# Twin Configuration
twin:
personality: "professional" # professional, friendly, technical
learning_rate: 0.8 # How quickly Twin learns from interactions
context_retention: 365 # Days of context to retain
# Schema Agent Configuration
schema_agent:
auto_evolve: true # Automatically evolve schemas
strict_mode: false # Strict validation vs lenient
# Coworker Configuration
coworker:
collaboration_mode: "active" # active, passive, disabled
task_timeout: 3600 # Seconds before task timeout
# Discovery Configuration
discovery:
search_depth: 3 # How deep to search for capabilities
cache_ttl: 300 # Cache time-to-live in seconds
# Triage Configuration
triage:
priority_thresholds: # Custom priority thresholds
urgent: 0.9
high: 0.7
normal: 0.5
low: 0.3
Monitoring
Core Agent Metrics
| Agent | Metrics | Description |
|---|---|---|
| Twin | Requests Processed | Total user requests handled |
| Context Assembled | Context packages created | |
| Agent Coordinated | Number of agents coordinated | |
| Schema | Schemas Validated | Data validation operations |
| Schemas Evolved | Automatic schema updates | |
| Mappings Performed | Entity mapping operations | |
| Coworker | Tasks Collaborated | Shared tasks completed |
| Handoffs Managed | Work transitions handled | |
| Discovery | Capabilities Found | Successful capability discoveries |
| Tools Discovered | Tool discovery operations | |
| Triage | Requests Routed | Work routing operations |
| Priority Decisions | Priority assessments made |
Health Checks
All Core Agents include:
- Heartbeat - Regular health signals
- Performance Metrics - Response times and throughput
- Error Tracking - Error rates and types
- Resource Usage - CPU, memory, and network usage
Troubleshooting
Common Issues
| Issue | Cause | Solution |
|---|---|---|
| Slow response times | High agent load | Check load balancing, scale resources |
| Context missing | Incomplete data | Verify integrations, check permissions |
| Wrong agent selected | Discovery misconfiguration | Review capability definitions, retrain discovery |
| Work not routed | Triage rules incorrect | Update priority thresholds, verify agent availability |
Debug Mode
Enable debug mode for detailed logging:
# Enable debug for specific agent
integratewise agent debug twin --level verbose
# View agent logs
integratewise agent logs twin --tail 100
# Check agent health
integratewise agent health --all
Best Practices
For Administrators
- Monitor Core Agents - Keep an eye on core agent health and performance
- Review Discovery Index - Regularly update capability and tool registries
- Tune Triage Rules - Adjust priority thresholds based on your organization's needs
- Backup Schema - Regularly backup your schema definitions
For Developers
- Use Discovery API - Leverage Discovery Agent to find capabilities programmatically
- Register Capabilities - Ensure all new agents and tools are properly registered
- Follow Schema - Always validate data against canonical schemas
- Handle Handoffs - Design agents to properly receive and pass context
For Users
- Provide Feedback - Help Twin learn your preferences
- Be Specific - Clear requests help agents work better
- Review Context - Check the context Twin assembles for accuracy
- Use Approvals - Configure approval workflows for sensitive actions
Roadmap
Core Agent Enhancements
- Twin - Improved personalization and learning algorithms
- Schema - Enhanced schema inference and conflict resolution
- Coworker - Better multi-human collaboration support
- Discovery - Advanced capability matching and ranking
- Triage - Predictive routing and load optimization
Support
Documentation
- Spine Agent Network - Overview of the agent network
- Domain Agents - Specialized agents by function
- Network Effect - How agents create exponential value
Community
Quick Reference
Core Agent Summary
| Agent | Nickname | Primary Role | Always Active |
|---|---|---|---|
| Twin | "The Boss" | Orchestration | ✅ Yes |
| Schema | "The Librarian" | Data Architecture | ✅ Yes |
| Coworker | "The Sidekick" | Collaboration | ✅ Yes |
| Discovery | "The Scout" | Information Finding | ✅ Yes |
| Triage | "The Traffic Cop" | Work Routing | ✅ Yes |
Key Concepts
- Twin-Centric - All interactions go through Twin
- Context-First - Context flows with every request
- Handoff - Work transfer between agents with full context
- Governed - All actions have proper permissions and audit