AgentOS Examples

Basic Agent

Beginner

Create and run your first AI agent with a simple task.

agent.yaml

name: my_first_agent
model_provider: github
model_version: openai/gpt-4o-mini
isolated: false

bash

agentos run agent.yaml --task "create a Python script that prints 'Hello, World!'"

✓ Agent 'my_first_agent' started
✓ Task received: create a Python script that prints 'Hello, World!'
✓ Creating file: hello.py

# Generated hello.py:
print("Hello, World!")

✓ Task completed successfully

Interactive Chat

Beginner

Start an interactive conversation with different LLM providers.

bash

# Chat with GitHub Models (default)
agentos chat --provider github

# Requires: GIT_HUB_TOKEN environment variable

bash

# Chat with Claude 3.5
agentos chat --provider claude --temperature 0.5

# Requires: CLAUDE_API_KEY environment variable

bash

# Chat with local Ollama (no API key needed!)
agentos chat --provider ollama

# Make sure Ollama is running locally
# Install from: https://ollama.ai

Code Generator Agent

Intermediate

An agent specialized in generating Python code with custom restrictions.

code_generator.yaml

name: code_generator
model_provider: openai
model_version: gpt-4
isolated: false

DESTRUCTIVE_COMMANDS:
  - rm
  - rmdir
  - sudo
  - chmod
  - chown
  - kill

bash

# Generate a REST API
agentos run code_generator.yaml --task "create a Flask REST API with CRUD operations for a todo list"

# Generate a CLI tool
agentos run code_generator.yaml --task "create a Python CLI tool for converting CSV to JSON"

# Generate tests
agentos run code_generator.yaml --task "create pytest unit tests for the todo API"

Research Agent

Intermediate

An agent for research and analysis tasks with Claude's advanced reasoning.

research_agent.yaml

name: research_assistant
model_provider: claude
model_version: claude-3-5-sonnet
isolated: true

DESTRUCTIVE_COMMANDS:
  - rm
  - sudo
  - rmdir
  - dd
  - mkfs

bash

# Research and summarize
agentos run research_agent.yaml --task "analyze the Python files in ./src and create a summary of the code architecture"

# Data analysis
agentos run research_agent.yaml --task "read data.csv and create a statistical analysis report"

Secure Docker Agent

Advanced

Maximum security configuration with Docker isolation enabled.

secure_agent.yaml

name: secure_agent
model_provider: github
model_version: openai/gpt-4o-mini
isolated: true

DESTRUCTIVE_COMMANDS:
  - rm
  - rmdir
  - sudo
  - su
  - dd
  - mkfs
  - fdisk
  - format
  - kill
  - killall
  - pkill
  - chmod
  - chown
  - wget
  - curl

bash

# Ensure Docker is running
docker ps

# Run secure agent
agentos run secure_agent.yaml --task "process untrusted input safely" --timeout 60

Docker Required

When isolated: true is set, Docker must be running on your system.

Multi-Provider Workflow

Advanced

Use different providers for different tasks in a workflow.

workflow.sh

#!/bin/bash

# Step 1: Generate code with GPT-4
agentos run gpt4_agent.yaml --task "create a Python web scraper for news headlines"

# Step 2: Review code with Claude
agentos run claude_agent.yaml --task "review scraper.py for security issues and best practices"

# Step 3: Generate tests with Gemini
agentos run gemini_agent.yaml --task "create comprehensive tests for scraper.py"

echo "Workflow complete!"

Examples

Basic Agent

Interactive Chat

Code Generator Agent

Research Agent

Secure Docker Agent

Multi-Provider Workflow

Common Use Cases

Code Generation

Debugging

Documentation

Testing

Code Review

Automation

Try these examples yourself