@baselineos/autonomy

The agent operations manager for Baseline Protocol. Controls how far AI can act independently, when to escalate, and how trust is earned and lost. Manages the full agent lifecycle (deploy, scale, retire), maintains a capability registry, enforces safety protocols with violation detection, and integrates with the Mastra engine for multi-agent coordination.

Primary Exports

| Export | Kind | Description | |---|---|---| | BaselineAutonomySystem | class | Top-level facade -- agent management, system status, shutdown | | AutonomousAgent | class | Individual agent with task execution and safety checks | | AgentLifecycleManager | class | Deploy, scale, monitor, and retire agents across stages | | BaselineAutonomyIntegration | class | Migration and comparison tooling for existing agent systems | | MastraAutonomyEngine | class | Mastra-based engine for multi-agent reasoning and coordination | | MastraAutonomousAgent | class | Mastra-native agent with reasoning steps and safety monitoring |

Key Types

AgentConfig, AgentTask, TaskResult, SafetyProtocol, SafetyAlert, PerformanceMetrics, ScalingPolicy, CoordinationPlan

Usage

import { BaselineAutonomySystem } from '@baselineos/autonomy';

const autonomy = new BaselineAutonomySystem();
const status = autonomy.getSystemStatus();
// { status: '...', agents: [...] }

// Clean shutdown
await autonomy.shutdown();

Integration

• Depends on: @baselineos/protocol-core
• Consumed by: @baselineos/cli, baselineos

pnpm --filter @baselineos/autonomy build
pnpm --filter @baselineos/autonomy test