Executes and authors end-to-end tests for Ark agentic resources using the Chainsaw testing framework.
This skill provides a specialized environment for running and writing comprehensive E2E tests for agentic applications built on Ark. It enables developers to execute test suites, debug failures using advanced flags like resource retention, and create new test scenarios using standardized manifest patterns. By supporting both mock-llm configurations for cost-effective local testing and real LLM providers like Azure OpenAI for production validation, it ensures that agents, models, and queries behave as expected across the entire lifecycle.
Key Features
01Standardized test structure for models, agents, and queries
02Environment variable management for Azure OpenAI integrations
03Debug failures using resource retention and pause-on-failure modes
04Support for mock-llm configurations to reduce testing costs
056 GitHub stars
06Execute standard and specific E2E test suites with simple commands
Use Cases
01Building automated CI/CD pipelines for agentic software using mock simulations
02Validating new agentic resource configurations before deployment
03Debugging complex agent interaction failures in a controlled environment
