Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

README.md

AgentProbe Examples

Comprehensive examples to get you started with AgentProbe testing.

📋 Table of Contents

Category Example Description
Basic hello-world.yaml Simplest possible test — verify an agent responds
Basic multi-assertion.yaml Multiple assertions on a single response
Tools mock-tools.yaml Mock tool responses for deterministic testing
Tools fault-injection.yaml Inject tool failures to test error handling
Security prompt-injection.yaml Test resistance to prompt injection attacks
Security data-exfil.yaml Detect data exfiltration attempts
Contracts support-agent.yaml Contract testing for a customer support agent
Chaos tool-failures.yaml Chaos testing with random tool failures
Multi-Agent handoff.yaml Multi-agent orchestration and handoff
Compliance gdpr-audit.yaml GDPR compliance verification
CI github-actions.yml GitHub Actions workflow integration
CI gitlab-ci.yml GitLab CI pipeline integration
CI pre-commit-hook.sh Git pre-commit hook for agent tests
Security clawguard-integration.yaml ClawGuard LLM-as-Judge security scanning
Multi-Agent crewai-test.yaml CrewAI crew collaboration testing
Multi-Agent autogen-test.yaml AutoGen group chat testing
Adapters openai.yaml OpenAI adapter configuration
Adapters anthropic.yaml Anthropic adapter configuration
MCP mcp-security-scan.yaml MCP server security scanning

🚀 Quick Start

# Run a single example
npx agentprobe run examples/basic/hello-world.yaml

# Run all examples in a category
npx agentprobe run examples/security/

# Run all examples
npx agentprobe run examples/

📖 Writing Your Own Tests

Each YAML file follows the AgentProbe test format:

name: "Test Name"
description: "What this test verifies"
agent:
  adapter: openai          # or anthropic, custom, etc.
  model: gpt-4o-mini
steps:
  - send: "Your prompt"
    assert:
      - type: contains
        value: "expected output"

See individual examples for advanced patterns like tool mocking, fault injection, and multi-agent testing.