skills.rest
Motium-AI
Motium-AIOFFICIAL
Explore all skills by this author →
GitHub

harness-test

Run Claude Code harness tests in an isolated sandbox.

Updated Aug 27, 2026
Search Tags:#automation#workflow#testing#Claude#hooks#sandbox#harness
One-click install
npx skills add https://github.com/Motium-AI/claude-code-toolkit --skill harness-test
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: harness-test
Source: https://github.com/Motium-AI/claude-code-toolkit/tree/main/config/skills/harness-test
Command: npx skills add https://github.com/Motium-AI/claude-code-toolkit --skill harness-test

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires jq, python3, openssl, timeout, claude, tmux, git, and includes scripts (resource) components.

What problem does it solve?

This Skill provides an isolated sandbox to validate Claude Code harness changes (hooks, skills, and settings) before integrating them into the main toolkit. It ensures modifications are tested in a controlled environment, reducing risk to the production workflow.

Core Features & Use Cases

  • Isolated sandbox that mirrors project harness configuration (config/hooks, config/skills, config/commands) with a separate settings.json
  • Automated harness validation: propagate uncommitted changes, spawn a fresh Claude session, run test prompts, and report results
  • Works with test cases under test-cases/ to verify hook and skill behaviors in a repeatable, automated manner

Quick Start

  1. Trigger tests with /harness-test in Claude to run harness validations against modified harness files
  2. Or run setup-harness-sandbox.sh to create an isolated sandbox, propagate changes, and execute tests with run-all-tests.sh
  3. Review the results in .claude/harness-test-state.json and the per-test logs for details

Frequently Asked Questions about harness-test

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

FAQPage Schema
How do I test Claude Code hooks and skills safely before deployment?▼

Testing harness changes in an isolated sandbox involves propagating uncommitted modifications to a separate environment, spawning a fresh Claude session, and running predefined test prompts. This method ensures hooks, skills, and settings modifications are validated reproducibly before integration.

Can I validate modified settings.json files without affecting my main toolkit?▼

Yes, you can validate modified settings.json files safely by running them within an isolated sandbox environment. This sandbox uses a separate configuration setup to execute test cases and report results, ensuring your main toolkit remains unaffected by uncommitted changes.

What dependencies are required to run automated harness validation workflows?▼

To run automated harness validation workflows, you need dependencies such as jq, python3, openssl, timeout, claude, tmux, and git. These tools support the sandbox setup, change propagation, test execution, and state reporting processes required for comprehensive validation.

How do I automate test case execution for Claude Code workflow changes?▼

Automating test case execution for Claude Code workflow changes involves placing tests in a dedicated directory and triggering an automated validation script. The system propagates changes to a sandbox, runs the tests, and generates per-test logs and a state file for result analysis.

Why should I use an isolated sandbox for testing hooks and skills?▼

Using an isolated sandbox for testing hooks and skills reduces risk to the production environment by providing a controlled setting that mirrors project configurations. It ensures that modifications are validated reproducibly across environments before being integrated into the main workflow.

Related Skills

Explore More Software Engineering Skills→
dislerdisler
379

Agent Sandboxes

Execute isolated code and tasks inside E2B sandboxes.

Community
Advanced
paulinghampaulingham
1

harness-config

Delegate non-markdown Claude Code harness changes to an infrastructure-engineer with worktree isolation.

Community
Intermediate
alangebalangeb

background

Manage asynchronous background processes and tmux sessions for concurrent task execution.

Community
Advanced
openkaidenopenkaiden
3

debug-image

Inspect container images and replicate integration test failures via shell access.

Official
Intermediate