What is Chainsaw in the context of Ark?

Chainsaw is a declarative testing tool used to perform end-to-end validation of Ark resources, ensuring that agents, models, and queries interact correctly within your infrastructure.

How do I debug a failing test?

You can use the provided debug commands like --skip-delete and --pause-on-failure. This keeps the test resources active so you can manually inspect the state of your cluster.

Can I test without incurring LLM costs?

Yes, this skill provides support for mock-llm-values.yaml configurations, allowing you to simulate LLM behavior and run tests locally or in CI without calling external APIs.

What is the standard test directory structure?

A standard test includes a chainsaw-test.yaml definition, a README.md, and a manifests folder containing YAML files for models, agents, and queries in a specific numerical sequence.

Ark Chainsaw Testing

Name: Ark Chainsaw Testing
Author: dwmkerr

bydwmkerr

•

セキュリティとテスト

Executes and scaffolds end-to-end Ark resource tests using the Chainsaw framework and mock-LLM configurations.

概要

The Ark Chainsaw Testing skill equips Claude Code with the specialized knowledge to manage the full lifecycle of E2E tests for Ark resources. It enables developers to run selective test suites, debug complex failures by persisting cluster resources, and generate standardized test structures including manifests for models, agents, and queries. By supporting both mock-LLM environments and real Azure OpenAI integrations, this skill ensures robust validation of cloud-native AI infrastructure while optimizing for developer productivity and cost-effective testing.

主な機能

Execute comprehensive or selective Chainsaw E2E tests using selectors
Automated scaffolding for test directories and YAML manifest patterns
Mock-LLM integration for testing without consuming API tokens
Advanced debugging with resource persistence and pause-on-failure modes
2 GitHub stars
Standardized patterns for Agent-Model-Query interaction validation

ユースケース

Debugging failed E2E tests by inspecting live resources in a cluster
Validating Ark resource configurations and Agent logic in CI/CD pipelines
Rapidly creating new regression tests for cloud-native AI workflows

概要

主な機能

Execute comprehensive or selective Chainsaw E2E tests using selectors
Automated scaffolding for test directories and YAML manifest patterns
Mock-LLM integration for testing without consuming API tokens
Advanced debugging with resource persistence and pause-on-failure modes
2 GitHub stars
Standardized patterns for Agent-Model-Query interaction validation

ユースケース

Debugging failed E2E tests by inspecting live resources in a cluster
Validating Ark resource configurations and Agent logic in CI/CD pipelines
Rapidly creating new regression tests for cloud-native AI workflows