General

Sales IQ Data Agent

A specialized AI agent built with LangGraph and FastMCP that focuses exclusively on business intelligence and data analysis. This agent uses Model Context Protocol (MCP) to connect to data sources and provide intelligent responses to business queries.

promptBeginner5 min to valuemarkdown

0 views

Jan 14, 2026

Prompt Playground

1 Variables

Fill Variables

MODEL_HOSTNAME>

Preview

# [MODEL_HOSTNAME>]ales IQ [MODEL_HOSTNAME>]ata [MODEL_HOSTNAME>]gent

[MODEL_HOSTNAME>] specialized [MODEL_HOSTNAME>]I agent built with [MODEL_HOSTNAME>]angGraph and Fast[MODEL_HOSTNAME>]CP that focuses exclusively on business intelligence and data analysis. [MODEL_HOSTNAME>]his agent uses [MODEL_HOSTNAME>]odel Context Protocol ([MODEL_HOSTNAME>]CP) to connect to data sources and provide intelligent responses to business queries.

## [MODEL_HOSTNAME>]verview

[MODEL_HOSTNAME>]he [MODEL_HOSTNAME>]ales IQ [MODEL_HOSTNAME>]ata [MODEL_HOSTNAME>]gent is a focused, production-ready [MODEL_HOSTNAME>]I assistant that:
- **[MODEL_HOSTNAME>]pecializes in data analysis** - [MODEL_HOSTNAME>]nly handles business intelligence queries
- **Uses [MODEL_HOSTNAME>]CP Protocol** - Connects to business data through standardized [MODEL_HOSTNAME>]CP servers
- **Built with [MODEL_HOSTNAME>]angGraph** - Provides robust workflow orchestration and tool routing
- **Fast[MODEL_HOSTNAME>]CP Integration** - [MODEL_HOSTNAME>]everages the Fast[MODEL_HOSTNAME>]CP library for reliable [MODEL_HOSTNAME>]CP connectivity

## Key Features

### 🎯 **[MODEL_HOSTNAME>]pecialized [MODEL_HOSTNAME>]ata Focus**
- Pure data analysis agent - no general-purpose tools
- Intelligent query routing to appropriate [MODEL_HOSTNAME>]CP tools
- Business intelligence and metrics generation
- [MODEL_HOSTNAME>]Q[MODEL_HOSTNAME>] query generation and execution

### 🔗 **[MODEL_HOSTNAME>]CP Protocol Integration**
- **Business Intelligence Query** - [MODEL_HOSTNAME>]atural language to [MODEL_HOSTNAME>]Q[MODEL_HOSTNAME>] conversion
- **Business [MODEL_HOSTNAME>]omain [MODEL_HOSTNAME>]iscovery** - [MODEL_HOSTNAME>]ist available data domains
- **[MODEL_HOSTNAME>]etric Creation** - Create custom business metrics from descriptions
- **[MODEL_HOSTNAME>]erver [MODEL_HOSTNAME>]anagement** - [MODEL_HOSTNAME>]ist tools and get server information

### 🤖 **[MODEL_HOSTNAME>]dvanced [MODEL_HOSTNAME>]gent Capabilities**
- **Retry [MODEL_HOSTNAME>]ogic** - [MODEL_HOSTNAME>]utomatically retries failed queries up to 3 attempts
- **Query Breakdown** - [MODEL_HOSTNAME>]plits complex requests into manageable parts
- **[MODEL_HOSTNAME>]Q[MODEL_HOSTNAME>] [MODEL_HOSTNAME>]xtraction** - [MODEL_HOSTNAME>]hows the actual [MODEL_HOSTNAME>]Q[MODEL_HOSTNAME>] queries used for transparency
- **[MODEL_HOSTNAME>]rror [MODEL_HOSTNAME>]andling** - Graceful failure management with user feedback

## Quick [MODEL_HOSTNAME>]tart

### Prerequisites
- Python 3.11+ (required for Fast[MODEL_HOSTNAME>]CP)
- [MODEL_HOSTNAME>]pen[MODEL_HOSTNAME>]I [MODEL_HOSTNAME>]PI key
- Running [MODEL_HOSTNAME>]CP server with business data

### Installation

1. **Clone and setup:**
```bash
git clone <your-repo-url[MODEL_HOSTNAME>]
cd sales-iq-agent-basic
python -m venv .venv
source .venv/bin/activate  # [MODEL_HOSTNAME>]n Windows: .venv\[MODEL_HOSTNAME>]cripts\activate
pip install -r requirements.txt
```

2. **Configure environment:**
```bash
cp .env.example .env
# [MODEL_HOSTNAME>]dit .env and add your [MODEL_HOSTNAME>]pen[MODEL_HOSTNAME>]I [MODEL_HOSTNAME>]PI key:
# [MODEL_HOSTNAME>]P[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]I[MODEL_HOSTNAME>][MODEL_HOSTNAME>]PI[MODEL_HOSTNAME>]K[MODEL_HOSTNAME>]Y=your[MODEL_HOSTNAME>]openai[MODEL_HOSTNAME>]api[MODEL_HOSTNAME>]key[MODEL_HOSTNAME>]here
```

3. **Configure [MODEL_HOSTNAME>]CP server:**
```bash
cp mcp[MODEL_HOSTNAME>]servers[MODEL_HOSTNAME>]config.template.json mcp[MODEL_HOSTNAME>]servers[MODEL_HOSTNAME>]config.json
# [MODEL_HOSTNAME>]dit mcp[MODEL_HOSTNAME>]servers[MODEL_HOSTNAME>]config.json with your [MODEL_HOSTNAME>]CP server details
```

4. **Run the agent:**
```bash
python main.py
```

## Configuration

### [MODEL_HOSTNAME>]CP [MODEL_HOSTNAME>]erver Configuration

[MODEL_HOSTNAME>]dit `mcp[MODEL_HOSTNAME>]servers[MODEL_HOSTNAME>]config.json`:
```json
{
  "mcp[MODEL_HOSTNAME>]ervers": {
    "sales[MODEL_HOSTNAME>]iq": {
      "transport": "http",
      "url": "http://<[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]:8000/mcp",
      "headers": {
        "[MODEL_HOSTNAME>]ccept": "application/json, text/event-stream",
        "Content-[MODEL_HOSTNAME>]ype": "application/json",
        "x-model": "Y[MODEL_HOSTNAME>]UR[MODEL_HOSTNAME>]IQ[MODEL_HOSTNAME>][MODEL_HOSTNAME>]pace",
        "x-api-key": "Y[MODEL_HOSTNAME>]UR[MODEL_HOSTNAME>][MODEL_HOSTNAME>]CP[MODEL_HOSTNAME>][MODEL_HOSTNAME>]PI[MODEL_HOSTNAME>]K[MODEL_HOSTNAME>]Y[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]R[MODEL_HOSTNAME>]"
      },
      "timeout": 600
    }
  }
}
```

### [MODEL_HOSTNAME>]nvironment Variables

[MODEL_HOSTNAME>]et in `.env` file:
```bash
# Required
[MODEL_HOSTNAME>]P[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]I[MODEL_HOSTNAME>][MODEL_HOSTNAME>]PI[MODEL_HOSTNAME>]K[MODEL_HOSTNAME>]Y=your[MODEL_HOSTNAME>]openai[MODEL_HOSTNAME>]api[MODEL_HOSTNAME>]key[MODEL_HOSTNAME>]here

# [MODEL_HOSTNAME>]ptional (can override config file)
[MODEL_HOSTNAME>]CP[MODEL_HOSTNAME>][MODEL_HOSTNAME>]PI[MODEL_HOSTNAME>]K[MODEL_HOSTNAME>]Y=your[MODEL_HOSTNAME>]mcp[MODEL_HOSTNAME>]api[MODEL_HOSTNAME>]key[MODEL_HOSTNAME>]here
[MODEL_HOSTNAME>]CP[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]RV[MODEL_HOSTNAME>]R[MODEL_HOSTNAME>]UR[MODEL_HOSTNAME>]=http://localhost:8000/mcp
[MODEL_HOSTNAME>]CP[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]=[MODEL_HOSTNAME>]ales [MODEL_HOSTNAME>]pace
```

## Usage [MODEL_HOSTNAME>]xamples

### Interactive [MODEL_HOSTNAME>]ode
```bash
python main.py
```

### [MODEL_HOSTNAME>]xample Queries
```
👤 You: What business domains are available for analysis?
🤖 [MODEL_HOSTNAME>]gent: [MODEL_HOSTNAME>]he available business domain is "[MODEL_HOSTNAME>]3C[MODEL_HOSTNAME>]PY[MODEL_HOSTNAME>][MODEL_HOSTNAME>]PC[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]F1"

👤 You: [MODEL_HOSTNAME>]how me total sales for Q4 1998
🤖 [MODEL_HOSTNAME>]gent: [MODEL_HOSTNAME>]he total sales for Q4 1998 was $1,468,808,711.04
         [MODEL_HOSTNAME>]Q[MODEL_HOSTNAME>] Query: [MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]C[MODEL_HOSTNAME>] [MODEL_HOSTNAME>]U[MODEL_HOSTNAME>]([MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]X[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]PRIC[MODEL_HOSTNAME>] * (1 - [MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]I[MODEL_HOSTNAME>]C[MODEL_HOSTNAME>]U[MODEL_HOSTNAME>][MODEL_HOSTNAME>]))...

👤 You: Find products with low inventory levels
🤖 [MODEL_HOSTNAME>]gent: Found 20 products with critical inventory (1 unit each):
         - Product I[MODEL_HOSTNAME>] 124510: "navajo spring linen lemon burlywood"
         - Product I[MODEL_HOSTNAME>] 5319: "white almond firebrick bisque metallic"
         ...
```

### Programmatic Usage
```python
from agent import create[MODEL_HOSTNAME>]agent

# Create specialized data agent
agent = create[MODEL_HOSTNAME>]agent()

# Query business data
response = agent.chat("What are the top 5 customers by revenue?")
print(response)
```

## Project [MODEL_HOSTNAME>]tructure

```
sales-iq-agent-basic/
├── agent.py                       # [MODEL_HOSTNAME>]ain [MODEL_HOSTNAME>]angGraph agent implementation
├── tools.py                       # [MODEL_HOSTNAME>]CP tool definitions
├── mcp[MODEL_HOSTNAME>]client.py                  # Fast[MODEL_HOSTNAME>]CP client implementation
├── main.py                        # C[MODEL_HOSTNAME>]I application entry point
├── requirements.txt               # Python dependencies
├── .env.example                   # [MODEL_HOSTNAME>]nvironment variables template
├── .gitignore                     # Git ignore rules
├── mcp[MODEL_HOSTNAME>]servers[MODEL_HOSTNAME>]config.template.json # [MODEL_HOSTNAME>]CP configuration template
├── test[MODEL_HOSTNAME>]proper[MODEL_HOSTNAME>]mcp.py             # [MODEL_HOSTNAME>]CP integration tests
└── R[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>].md                      # [MODEL_HOSTNAME>]his documentation
```

## [MODEL_HOSTNAME>]rchitecture

### [MODEL_HOSTNAME>]gent Workflow
```
User Query → [MODEL_HOSTNAME>]ata [MODEL_HOSTNAME>]gent → [MODEL_HOSTNAME>]ecision → [MODEL_HOSTNAME>]CP [MODEL_HOSTNAME>]ools → [MODEL_HOSTNAME>]ata [MODEL_HOSTNAME>]gent → Response
                    ↓              ↑
                   [MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>]        ([MODEL_HOSTNAME>]ool Results)
```

### [MODEL_HOSTNAME>]CP [MODEL_HOSTNAME>]ools [MODEL_HOSTNAME>]vailable
1. **`business[MODEL_HOSTNAME>]intelligence[MODEL_HOSTNAME>]query`** - [MODEL_HOSTNAME>]atural language to [MODEL_HOSTNAME>]Q[MODEL_HOSTNAME>] with results
2. **`get[MODEL_HOSTNAME>]available[MODEL_HOSTNAME>]business[MODEL_HOSTNAME>]domains`** - [MODEL_HOSTNAME>]ist business data domains  
3. **`create[MODEL_HOSTNAME>]business[MODEL_HOSTNAME>]metric[MODEL_HOSTNAME>]tool`** - Create custom metrics
4. **`list[MODEL_HOSTNAME>]mcp[MODEL_HOSTNAME>]tools`** - [MODEL_HOSTNAME>]how available [MODEL_HOSTNAME>]CP server tools
5. **`get[MODEL_HOSTNAME>]mcp[MODEL_HOSTNAME>]server[MODEL_HOSTNAME>]info`** - [MODEL_HOSTNAME>]isplay server capabilities

### [MODEL_HOSTNAME>]rror [MODEL_HOSTNAME>]andling & Retries
- **3-attempt retry logic** for failed queries
- **[MODEL_HOSTNAME>]utomatic query breakdown** for complex requests
- **Graceful error messaging** with actionable feedback
- **[MODEL_HOSTNAME>]Q[MODEL_HOSTNAME>] query transparency** - shows actual queries used

## [MODEL_HOSTNAME>]evelopment

### [MODEL_HOSTNAME>]esting [MODEL_HOSTNAME>]CP Integration
```bash
python test[MODEL_HOSTNAME>]proper[MODEL_HOSTNAME>]mcp.py
```

### [MODEL_HOSTNAME>]esting [MODEL_HOSTNAME>]pecific [MODEL_HOSTNAME>]ools
```python
from mcp[MODEL_HOSTNAME>]client import business[MODEL_HOSTNAME>]intelligence[MODEL_HOSTNAME>]query

# [MODEL_HOSTNAME>]est business intelligence
result = business[MODEL_HOSTNAME>]intelligence[MODEL_HOSTNAME>]query("[MODEL_HOSTNAME>]how sales trends by quarter")
print(result)
```

### [MODEL_HOSTNAME>]isplaying [MODEL_HOSTNAME>]gent Workflow
```python
from agent import create[MODEL_HOSTNAME>]agent, display[MODEL_HOSTNAME>]graph

agent = create[MODEL_HOSTNAME>]agent()
display[MODEL_HOSTNAME>]graph(agent.graph)  # Creates agent[MODEL_HOSTNAME>]workflow.png
```

## [MODEL_HOSTNAME>]ependencies

### Core [MODEL_HOSTNAME>]ibraries
- **[MODEL_HOSTNAME>]angGraph** - [MODEL_HOSTNAME>]gent workflow orchestration
- **[MODEL_HOSTNAME>]angChain** - [MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>] integration and tool management
- **Fast[MODEL_HOSTNAME>]CP** - [MODEL_HOSTNAME>]odel Context Protocol client
- **[MODEL_HOSTNAME>]pen[MODEL_HOSTNAME>]I** - [MODEL_HOSTNAME>]arge language model provider

### Complete [MODEL_HOSTNAME>]ependencies
[MODEL_HOSTNAME>]ee `requirements.txt` for full list with versions.

## [MODEL_HOSTNAME>]roubleshooting

### Common Issues

**1. "[MODEL_HOSTNAME>]o [MODEL_HOSTNAME>]CP tools available"**
- Check `mcp[MODEL_HOSTNAME>]servers[MODEL_HOSTNAME>]config.json` exists and has valid configuration
- Verify [MODEL_HOSTNAME>]CP server is running and accessible
- Confirm [MODEL_HOSTNAME>]PI key is correct

**2. "Invalid [MODEL_HOSTNAME>]PI Key" errors**
- Verify [MODEL_HOSTNAME>]pen[MODEL_HOSTNAME>]I [MODEL_HOSTNAME>]PI key in `.env` file
- Check [MODEL_HOSTNAME>]CP [MODEL_HOSTNAME>]PI key in config file
- [MODEL_HOSTNAME>]nsure [MODEL_HOSTNAME>]PI keys have proper permissions

**3. Connection timeouts**
- Increase timeout in `mcp[MODEL_HOSTNAME>]servers[MODEL_HOSTNAME>]config.json`
- Check network connectivity to [MODEL_HOSTNAME>]CP server
- Verify server UR[MODEL_HOSTNAME>] and port

**4. "[MODEL_HOSTNAME>]ool call failed" errors**
- Check [MODEL_HOSTNAME>]CP server logs for detailed errors
- Verify data source connectivity
- [MODEL_HOSTNAME>]est with simpler queries first

### [MODEL_HOSTNAME>]ebug [MODEL_HOSTNAME>]ode
```bash
python main.py --no-api-key-check  # [MODEL_HOSTNAME>]kip [MODEL_HOSTNAME>]PI validation
```

## Contributing

[MODEL_HOSTNAME>]his is a specialized data agent focusing on business intelligence through [MODEL_HOSTNAME>]CP. Contributions should maintain this focus:

1. Keep tools limited to data analysis and [MODEL_HOSTNAME>]CP integration
2. [MODEL_HOSTNAME>]aintain error handling and retry logic
3. [MODEL_HOSTNAME>]nsure security best practices (no hardcoded keys)
4. Update tests for any new [MODEL_HOSTNAME>]CP tools

## [MODEL_HOSTNAME>]icense

[MODEL_HOSTNAME>]I[MODEL_HOSTNAME>] [MODEL_HOSTNAME>]icense - see [MODEL_HOSTNAME>]IC[MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>][MODEL_HOSTNAME>] file for details.

## [MODEL_HOSTNAME>]upport

For issues:
1. Check the troubleshooting section
2. Verify [MODEL_HOSTNAME>]CP server configuration
3. [MODEL_HOSTNAME>]est [MODEL_HOSTNAME>]CP connectivity with `test[MODEL_HOSTNAME>]proper[MODEL_HOSTNAME>]mcp.py`
4. Check agent logs for detailed error information

Sales IQ Data Agent

Overview

The Sales IQ Data Agent is a focused, production-ready AI assistant that:

Specializes in data analysis - Only handles business intelligence queries
Uses MCP Protocol - Connects to business data through standardized MCP servers
Built with LangGraph - Provides robust workflow orchestration and tool routing
FastMCP Integration - Leverages the FastMCP library for reliable MCP connectivity

Key Features

🎯 Specialized Data Focus

Pure data analysis agent - no general-purpose tools
Intelligent query routing to appropriate MCP tools
Business intelligence and metrics generation
SQL query generation and execution

🔗 MCP Protocol Integration

Business Intelligence Query - Natural language to SQL conversion
Business Domain Discovery - List available data domains
Metric Creation - Create custom business metrics from descriptions
Server Management - List tools and get server information

🤖 Advanced Agent Capabilities

Retry Logic - Automatically retries failed queries up to 3 attempts
Query Breakdown - Splits complex requests into manageable parts
SQL Extraction - Shows the actual SQL queries used for transparency
Error Handling - Graceful failure management with user feedback

Quick Start

Prerequisites

Python 3.11+ (required for FastMCP)
OpenAI API key
Running MCP server with business data

Installation

Clone and setup:

git clone <your-repo-url>
cd sales-iq-agent-basic
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install -r requirements.txt

Configure environment:

cp .env.example .env
# Edit .env and add your OpenAI API key:
# OPENAI_API_KEY=your_openai_api_key_here

Configure MCP server:

cp mcp_servers_config.template.json mcp_servers_config.json
# Edit mcp_servers_config.json with your MCP server details

Run the agent:

python main.py

Configuration

MCP Server Configuration

Edit

mcp_servers_config.json

{
  "mcpServers": {
    "sales_iq": {
      "transport": "http",
      "url": "http://<MODEL_HOSTNAME>:8000/mcp",
      "headers": {
        "Accept": "application/json, text/event-stream",
        "Content-Type": "application/json",
        "x-model": "YOUR_IQ_Space",
        "x-api-key": "YOUR_MCP_API_KEY_HERE"
      },
      "timeout": 600
    }
  }
}

Environment Variables

Set in

.env

file:

# Required
OPENAI_API_KEY=your_openai_api_key_here

# Optional (can override config file)
MCP_API_KEY=your_mcp_api_key_here
MCP_SERVER_URL=http://localhost:8000/mcp
MCP_MODEL=Sales Space

Usage Examples

Interactive Mode

python main.py

Example Queries

👤 You: What business domains are available for analysis?
🤖 Agent: The available business domain is "T3COPY_TPCH_SF1"

👤 You: Show me total sales for Q4 1998
🤖 Agent: The total sales for Q4 1998 was $1,468,808,711.04
         SQL Query: SELECT SUM(L_EXTENDEDPRICE * (1 - L_DISCOUNT))...

👤 You: Find products with low inventory levels
🤖 Agent: Found 20 products with critical inventory (1 unit each):
         - Product ID 124510: "navajo spring linen lemon burlywood"
         - Product ID 5319: "white almond firebrick bisque metallic"
         ...

Programmatic Usage

from agent import create_agent

# Create specialized data agent
agent = create_agent()

# Query business data
response = agent.chat("What are the top 5 customers by revenue?")
print(response)

Project Structure

sales-iq-agent-basic/
├── agent.py                       # Main LangGraph agent implementation
├── tools.py                       # MCP tool definitions
├── mcp_client.py                  # FastMCP client implementation
├── main.py                        # CLI application entry point
├── requirements.txt               # Python dependencies
├── .env.example                   # Environment variables template
├── .gitignore                     # Git ignore rules
├── mcp_servers_config.template.json # MCP configuration template
├── test_proper_mcp.py             # MCP integration tests
└── README.md                      # This documentation

Architecture

Agent Workflow

User Query → Data Agent → Decision → MCP Tools → Data Agent → Response
                    ↓              ↑
                   END        (Tool Results)

MCP Tools Available

business_intelligence_query
- Natural language to SQL with results
get_available_business_domains
- List business data domains
create_business_metric_tool
- Create custom metrics
list_mcp_tools
- Show available MCP server tools
get_mcp_server_info
- Display server capabilities

Error Handling & Retries

3-attempt retry logic for failed queries
Automatic query breakdown for complex requests
Graceful error messaging with actionable feedback
SQL query transparency - shows actual queries used

Development

Testing MCP Integration

python test_proper_mcp.py

Testing Specific Tools

from mcp_client import business_intelligence_query

# Test business intelligence
result = business_intelligence_query("Show sales trends by quarter")
print(result)

Displaying Agent Workflow

from agent import create_agent, display_graph

agent = create_agent()
display_graph(agent.graph)  # Creates agent_workflow.png

Dependencies

Core Libraries

LangGraph - Agent workflow orchestration
LangChain - LLM integration and tool management
FastMCP - Model Context Protocol client
OpenAI - Large language model provider

Complete Dependencies

See

requirements.txt

for full list with versions.

Troubleshooting

Common Issues

1. "No MCP tools available"

Check
```
mcp_servers_config.json
```
exists and has valid configuration
Verify MCP server is running and accessible
Confirm API key is correct

2. "Invalid API Key" errors

Verify OpenAI API key in
```
.env
```
file
Check MCP API key in config file
Ensure API keys have proper permissions

3. Connection timeouts

Increase timeout in
```
mcp_servers_config.json
```
Check network connectivity to MCP server
Verify server URL and port

4. "Tool call failed" errors

Check MCP server logs for detailed errors
Verify data source connectivity
Test with simpler queries first

Debug Mode

python main.py --no-api-key-check  # Skip API validation

Contributing

This is a specialized data agent focusing on business intelligence through MCP. Contributions should maintain this focus:

Keep tools limited to data analysis and MCP integration
Maintain error handling and retry logic
Ensure security best practices (no hardcoded keys)
Update tests for any new MCP tools

License

MIT License - see LICENSE file for details.

Support

For issues:

Check the troubleshooting section
Verify MCP server configuration
Test MCP connectivity with
```
test_proper_mcp.py
```
Check agent logs for detailed error information

View Original Source

Related Skills

General

PromptBeginner5 minmarkdown

<h1 align="center">

Jan 12, 2026

General

PromptBeginner5 minmarkdown

- Identify gaps

ambiguities

Jan 15, 2026

General

PromptBeginner5 minmarkdown

2. Apply Deepthink Protocol (reason about dependencies

risks

Jan 15, 2026