Is there a visual dashboard for monitoring HarshJudge tests?

Yes, HarshJudge includes a dashboard command (`harshjudge dashboard`) to visually manage and view the status of your test runs.

What types of evidence does HarshJudge capture?

HarshJudge captures screenshots, console logs, and HTML for frontend tests; response bodies and headers for APIs; and stdout/stderr/exit codes for CLI tests.

How does HarshJudge optimize token usage during tests?

It spawns separate agents for individual steps and saves large outputs (like screenshots or logs) to files rather than returning them directly to the main agent's context.

What is the benefit of the step-based agent pattern in HarshJudge?

It ensures each test step runs in an isolated context, preventing token accumulation and making the testing process more efficient and reliable for long-running scenarios.

Does HarshJudge support both frontend and backend testing?

Yes, it is designed to orchestrate tests across frontend (browser), backend (API), and CLI environments seamlessly within a single framework.

HarshJudge E2E Testing Orchestrator

Name: HarshJudge E2E Testing Orchestrator
Author: HuskyDanny

byHuskyDanny

0•

Seguridad y Pruebas

Orchestrates comprehensive end-to-end testing across frontend, backend, and CLI environments using an AI-native step-based agent pattern.

HarshJudge is a specialized testing framework for Claude Code designed to manage complex E2E scenarios with maximum efficiency. It utilizes a unique step-based agent architecture where individual test steps are executed by spawned agents, preventing token accumulation and ensuring isolated execution contexts. The skill automatically handles evidence capture—including screenshots for web apps, response bodies for APIs, and stdout for CLI tools—storing them locally to keep the main agent's context clean. It is an essential tool for developers looking to automate regression testing, validate user journeys, and maintain high-quality codebases within the Claude Code environment.

Características Principales

01Step-based agent execution for isolated context and optimized token usage

02Cross-platform support for frontend (browser), backend (API), and CLI testing

03Automated evidence capture including screenshots, API logs, and CLI output

04Fail-fast orchestration that stops on errors with detailed context reporting

05Built-in CLI and dashboard for managing test runs and project status

060 GitHub stars

Casos de Uso

01Validating CLI tool behavior and infrastructure scripts with stdout and exit code checks

02Running regression suites against API endpoints with automated response validation

03Automating complex multi-step user journeys in web applications with visual verification

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add huskydanny/harshjudge harshjudge

For use in Claude.ai and ChatGPT