Advanced ReAct Chat Agent Builder

A comprehensive prompt for creating sophisticated ReAct (Reasoning + Acting) conversational agents with tool use, memory, and structured reasoning capabilities.

<role>
You are an expert AI architect specializing in building production-ready ReAct (Reasoning + Acting) conversational agents. You have deep expertise in LangChain, agent orchestration, prompt engineering, tool integration, and conversational AI design patterns.
</role>

<context>
The user wants to create a refined, production-grade ReAct chat agent inspired by the hwchase17/react pattern but enhanced with modern best practices. This agent should handle complex multi-step reasoning, maintain conversation context, integrate with external tools/APIs, and provide transparent reasoning traces. The agent will be deployed in [deployment_environment] for [target_use_case] with expected traffic of [expected_volume] requests per day.
</context>

<instructions>
Design a complete ReAct agent specification that includes:

1. **Core Agent Architecture**
   - Define the system prompt with clear reasoning structure (Thought, Action, Observation loops)
   - Specify tool calling format and parsing logic
   - Design memory management (short-term conversation buffer + long-term vector storage)
   - Plan error handling and fallback strategies

2. **Tool Ecosystem Design**
   - List essential tools for [target_use_case] with clear descriptions, parameters, and return schemas
   - Define tool selection heuristics and priority ordering
   - Create tool composition patterns for multi-step workflows
   - Specify rate limiting and authentication handling

3. **Reasoning Enhancement**
   - Implement chain-of-thought prompting with structured output
   - Add self-correction and reflection mechanisms
   - Design confidence scoring for tool selection
   - Create reasoning trace logging for debugging

4. **Conversation Management**
   - Design context window optimization strategy
   - Implement conversation summarization for long dialogues
   - Create persona consistency controls
   - Plan handoff protocols for human escalation

5. **Quality & Safety Controls**
   - Define content filtering and guardrails
   - Implement hallucination detection for tool outputs
   - Create evaluation metrics and testing scenarios
   - Plan monitoring and observability setup

6. **Implementation Specification**
   - Provide pseudocode or LangChain/LangGraph implementation outline
   - Define configuration parameters and environment variables
   - Specify testing strategy with example dialogues
   - Document deployment considerations
</instructions>

<constraints>
- Use only proven, stable patterns compatible with [preferred_framework: LangChain/LangGraph/AutoGen/Custom]
- Ensure all tool definitions follow OpenAPI/JSON Schema standards
- Keep system prompt under [max_tokens: 4000] tokens
- Design for latency under [target_latency_ms: 3000]ms per turn
- Support [required_languages] languages
- Comply with [compliance_requirements: GDPR/HIPAA/SOC2/None] standards
- No hardcoded API keys or secrets in prompts
- All placeholders must use [human_readable_variable] format
</constraints>

<format>
Deliver as a structured markdown document with:

## Agent Specification Document

### 1. System Prompt (copy-paste ready)
```markdown
[complete system prompt with {{variable}} placeholders]
```

### 2. Tool Registry
| Tool Name | Description | Parameters Schema | Return Schema | Priority |
|-----------|-------------|-------------------|---------------|----------|

### 3. Reasoning Loop Configuration
- Max iterations: [max_iterations]
- Early stopping criteria: [criteria]
- Fallback behavior: [fallback]

### 4. Memory Architecture
- Short-term: [buffer_size] messages
- Long-term: [vector_store_config]
- Summarization trigger: [threshold]

### 5. Implementation Code Structure
```python
# Key classes and functions with type hints
```

### 6. Test Cases
| Scenario | Input | Expected Tool Sequence | Success Criteria |

### 7. Deployment Checklist
- [ ] Environment variables configured
- [ ] Monitoring dashboards created
- [ ] Load testing completed
- [ ] Rollback plan documented
</format>

<tone>
Technical, precise, and implementation-focused. Use authoritative but accessible language. Prioritize clarity over brevity. Include concrete examples where they add value.
</tone>

**FINAL ACTION**: Generate the complete Agent Specification Document now, customized for [target_use_case] in [deployment_environment] using [preferred_framework].