agent-workflow-patterns

# Agent Workflow Patterns

**Purpose:** Provide production-ready agent architectures, workflow patterns, and loop control strategies for building autonomous AI systems with Vercel AI SDK.

**Activation Triggers:**
- Building autonomous AI agents
- Implementing multi-step reasoning
- Creating agent workflows
- Tool orchestration and coordination
- Loop control and iteration management
- Multi-agent system architectures
- ReAct (Reasoning + Acting) patterns

**Key Resources:**
- `templates/react-agent.ts` - ReAct agent pattern
- `templates/multi-agent-system.ts` - Multiple specialized agents
- `templates/workflow-orchestrator.ts` - Workflow coordination
- `templates/loop-control.ts` - Iteration and safeguards
- `templates/tool-coordinator.ts` - Tool orchestration
- `scripts/validate-agent.sh` - Validate agent configuration
- `examples/` - Production agent implementations (RAG agent, SQL agent, etc.)

## Core Agent Patterns

### 1. ReAct Agent (Reasoning + Acting)

**When to use:** Complex problem-solving requiring iterative thought and action

**Template:** `templates/react-agent.ts`

**Pattern:**
```typescript
async function reactAgent(task: string, maxIterations: number = 5) {
  const tools = { /* tool definitions */ }
  let iteration = 0

  while (iteration < maxIterations) {
    // Reasoning step
    const thought = await generateText({
      model: openai('gpt-4o')
      messages: [
        { role: 'system', content: 'Think step-by-step...' }
        { role: 'user', content: task }
      ]
    })

    // Acting step (tool calls)
    const action = await generateText({
      model: openai('gpt-4o')
      tools
      toolChoice: 'auto'
      messages: [/* ... */]
    })

    // Check if task complete
    if (isComplete(action)) break
    iteration++
  }

  return result
}
```

**Best for:** Research, analysis, complex planning

### 2. Multi-Agent System

**When to use:** Complex domains requiring specialized expertise

**Template:** `templates/multi-agent-system.ts