SCBE Fleet Management System
Secure AI Agent Fleet Orchestration with Sacred Tongue Governance
Overview
The Fleet Management System integrates SCBE’s security framework with AI workflow orchestration, enabling secure management of AI agent fleets with trust-based task assignment and Sacred Tongue governance.
Architecture
┌─────────────────────────────────────────────────────────────────┐
│ FleetManager │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────────────────┐ │
│ │ Agent │ │ Task │ │ Governance │ │
│ │ Registry │ │ Dispatcher │ │ Manager │ │
│ └──────┬──────┘ └──────┬──────┘ └───────────┬─────────────┘ │
│ │ │ │ │
│ ┌──────┴──────┐ ┌──────┴──────┐ ┌───────────┴─────────────┐ │
│ │ Trust │ │ Spectral │ │ Roundtable │ │
│ │ Manager │ │ Identity │ │ Consensus │ │
│ └─────────────┘ └─────────────┘ └─────────────────────────┘ │
└─────────────────────────────────────────────────────────────────┘
Quick Start
import { FleetManager, createDefaultFleet } from 'scbe-aethermoore/fleet';
// Create fleet with default agents
const fleet = createDefaultFleet();
// Or create custom fleet
const customFleet = new FleetManager({ autoAssign: true });
// Register an agent
const agent = customFleet.registerAgent({
name: 'CodeGen-GPT4',
description: 'Code generation specialist',
provider: 'openai',
model: 'gpt-4o',
capabilities: ['code_generation', 'code_review'],
maxGovernanceTier: 'CA',
initialTrustVector: [0.7, 0.6, 0.8, 0.5, 0.6, 0.4]
});
// Create and assign a task
const task = customFleet.createTask({
name: 'Generate API Handler',
description: 'Create REST API endpoint',
requiredCapability: 'code_generation',
requiredTier: 'RU',
priority: 'high',
input: { prompt: 'Create a user authentication endpoint' }
});
// Complete the task
customFleet.completeTask(task.id, { code: '...' });
Sacred Tongue Governance Tiers
| Tier | Name | Min Trust | Required Tongues | Operations |
|---|---|---|---|---|
| KO | Koraelin | 0.1 | 1 | Read-only |
| AV | Avali | 0.3 | 2 | Write |
| RU | Runethic | 0.5 | 3 | Execute |
| CA | Cassisivadan | 0.7 | 4 | Deploy |
| UM | Umbroth | 0.85 | 5 | Admin |
| DR | Draumric | 0.95 | 6 | Critical/Destructive |
Spectral Identity
Each agent receives a unique chromatic fingerprint based on their 6D trust vector:
// Agent's spectral identity
{
spectralHash: 'SP-A3F2-8B1C',
hexCode: '#4A7B9C',
colorName: 'Deep Blue',
confidence: 'HIGH'
}
Roundtable Consensus
Critical operations (UM/DR tier) require roundtable approval:
// Create roundtable for critical operation
const session = fleet.createRoundtable({
topic: 'Delete production database',
requiredTier: 'DR'
});
// Agents vote
fleet.castVote(session.id, agent1.id, 'approve');
fleet.castVote(session.id, agent2.id, 'approve');
// ... requires 6 approvals for DR tier
Event System
fleet.onEvent(event => {
switch (event.type) {
case 'agent_registered':
case 'task_completed':
case 'security_alert':
console.log(event);
}
});
Health Monitoring
const health = fleet.getHealthStatus();
// { healthy: true, issues: [], metrics: { ... } }
const stats = fleet.getStatistics();
// { totalAgents, tasksByStatus, fleetSuccessRate, ... }
Integration with AI Workflow Architect
The Fleet Management System is designed to integrate with the AI Workflow Architect’s:
- MCP server for tool execution
- Agent orchestration queue
- Roundtable sessions for multi-AI collaboration
- Audit logging for compliance
See external_repos/ai-workflow-architect/ for the full platform.