z007 0.2.1

Micro agent with tool support and MCP integration.

⚡ z007 🤖: Nimble AI Agent

pronounced: "zee-double-oh-seven"

A lightweight and readable agent for interacting with LLM on AWS Bedrock with tool and MCP (Model Context Protocol) support.

Features

• 🟢 Ultra Readable: Clean, maintainable codebase in about 600 lines - easy to understand, modify, and extend
• ⚡ Super easy: Just run uvx z007@latest with AWS_PROFILE=<your profile> in env and start chatting instantly
• ⚡ Simple Install: Quick install uv tool install --upgrade z007 and start chatting instantly z007 with AWS_PROFILE=<your profile> in env
• 🔧 Tool Support: Built-in calculator and easily use plain python functions as tools
• 🔌 MCP Integration: Connect to Model Context Protocol servers
• 🐍 Python API: Easy integration into your Python projects
• 🚀 Async: Concurrent tool execution

Quick Start

Install and run with uvx (recommended)

```bash
# Install and run directly with AWS_PROFILE configured - fastest way to start!
AWS_PROFILE=your-profile uvx z007@latest

# Or install globally
uv tool install z007
AWS_PROFILE=your-profile z007

Install as Python package

pip install z007

Usage

Command Line

# Start interactive chat
z007

# With custom model (AWS Bedrock)
AWS_PROFILE=your-profile z007 --model-id "openai.gpt-oss-120b-1:0"

# With MCP configuration
z007 --mcp-config ./mcp.json

Python API

Simple usage

import asyncio
from z007 import Agent, create_calculator_tool

async def main():
    calculator = create_calculator_tool()
    async with Agent(model_id="openai.gpt-oss-20b-1:0", tools=[calculator]) as agent:
        response = await agent.run("What is 2+2?")
    print(response)

asyncio.run(main())

Using the Agent class

import asyncio
from z007 import Agent, create_calculator_tool

async def main():
    calculator = create_calculator_tool()
    async with Agent(
        model_id="openai.gpt-oss-20b-1:0",
        system_prompt="You are a helpful coding assistant.",
        tools=[calculator]
    ) as agent:
        response = await agent.run("Write a Python function to reverse a string")
        print(response)

asyncio.run(main())

Custom Tools

Create your own tools by writing simple Python functions:

import asyncio
from z007 import Agent

def weather_tool(city: str) -> str:
    """Get weather information for a city"""
    # In a real implementation, call a weather API
    return f"The weather in {city} is sunny, 25°C"

def file_reader_tool(filename: str) -> str:
    """Read contents of a file"""
    try:
        with open(filename, 'r') as f:
            return f.read()
    except Exception as e:
        return f"Error reading file: {e}"

async def main():
    async with Agent(
        model_id="openai.gpt-oss-20b-1:0",
        tools=[weather_tool, file_reader_tool]
    ) as agent:
        response = await agent.run("What's the weather like in Paris?")
    print(response)

asyncio.run(main())

MCP Integration

Connect to Model Context Protocol servers for advanced capabilities:

1. Create mcp.json:

{
  "servers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/project"]
    },
    "brave-search": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-brave-search"],
      "env": {
        "BRAVE_API_KEY": "${env:BRAVE_API_KEY}"
      }
    },
    "playwright": {
      "command": "npx",
      "args": ["@playwright/mcp@latest"]
    }
  }
}

2. Use with z007:

z007 --mcp-config mcp.json

Or in Python:

import json
from z007 import Agent

# Load MCP config
with open("mcp.json") as f:
    mcp_config = json.load(f)

async with Agent(
    model_id="openai.gpt-oss-20b-1:0",
    mcp_config=mcp_config
) as agent:
    response = await agent.run("Search for recent news about AI")
    print(response)

Configuration

Environment Variables

For AWS Bedrock (default provider):

• AWS_PROFILE: AWS profile name (e.g., AWS_PROFILE=codemobs)

  or

• AWS_REGION: AWS region (default: us-east-1)

• AWS_ACCESS_KEY_ID: AWS access key

• AWS_SECRET_ACCESS_KEY: AWS secret key

Supported Models

AWS Bedrock models with verified access:

• openai.gpt-oss-20b-1:0 (default)

Note: Model availability depends on your AWS account's Bedrock access permissions. Use AWS_PROFILE=your-profile to specify credentials.

• Any AWS Bedrock model with tool support

Interactive Commands

When running z007 in interactive mode:

• /help - Show help
• /tools - List available tools
• /clear - Clear conversation history
• /exit - Exit

Requirements

• Python 3.9+
• LLM provider credentials (AWS for Bedrock)

License

MIT License