klayoutclaw:e2e-judge

Automate end-to-end agentic testing for KlayoutClaw with MCP and LLM judging.

32|6|Updated Mar 8, 2026

Search Tags:#automation #mcp #e2e-testing #judge #verifier #klayout

Ready to run

npx skills add https://github.com/caidish/KlayoutClaw --skill klayoutclaw-e2e-judge

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: klayoutclaw:e2e-judge
Source: https://github.com/caidish/KlayoutClaw/tree/main/skills/e2e_judge
Command: npx skills add https://github.com/caidish/KlayoutClaw --skill klayoutclaw-e2e-judge

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Orchestrates automated end-to-end testing of KlayoutClaw by running agentic tasks, performing independent MCP verifications, and judging outcomes with an LLM.

Core Features & Use Cases

Orchestrates a full E2E test pipeline: task generation, autonomous tool usage, layout verification, and judge verdicts.
Supports multiple test phases (preflight, layout, geometry, evaluate, hallbar, pipeline, discovery) and Phase 5 autonomous pipelines.
Provides an auditable transcript, verification results, and structured verdicts for performance benchmarking.

Quick Start

Launch the agentic E2E judge against a live KlayoutClaw MCP server and review the structured verdicts produced by the judge.

Frequently Asked Questions about klayoutclaw:e2e-judge

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate end-to-end testing for an MCP server using an LLM judge?▼

End-to-end testing for an MCP server using an LLM judge automates task generation, autonomous tool usage, layout verification, and verdicts. This Skill orchestrates the pipeline against a live KlayoutClaw MCP server to produce structured performance benchmarking results.

What do I need to run automated agentic E2E tests with KlayoutClaw?▼

To run automated agentic E2E tests, you need an MCP server running at 127.0.0.1:8765/mcp, Claude Code or qlaybot for agent execution, and the judge API endpoint. These dependencies enable the tool to orchestrate tasks and generate verdicts.

Can I use this E2E judge for full device design verification workflows?▼

Yes, you can use this E2E judge for full device design verification workflows. It supports multiple test phases including preflight, layout, geometry, evaluate, hallbar, pipeline, and discovery, spanning from reconnaissance to complete device design tests.

How does an LLM judge verify MCP tool usage in agentic testing?▼

An LLM judge verifies MCP tool usage by running agentic tasks, performing independent MCP verifications, and evaluating the outcomes. This process generates an auditable transcript, verification results, and structured verdicts for benchmarking.

Does this automated E2E testing pipeline support autonomous phases?▼

Yes, the automated E2E testing pipeline supports autonomous phases. It includes optional Phase 5 autonomous pipelines that run alongside the standard preflight, layout, geometry, evaluate, hallbar, pipeline, and discovery test phases.