Skill

test-userscript

Validate a userscript end-to-end in a real Chrome tab via the Claude in Chrome MCP. Loads the target page, simulates the script's effect, reads console logs and DOM state, and reports whether the script's behavior matches its `@description`.

npx claudepluginhub danielrosehill/claude-code-plugins --plugin userscript-development

Tool Access

This skill uses the workspace's default tool permissions.

Preview

Validate userscript behavior live in Chrome. Tampermonkey itself can't be installed/configured by Claude in Chrome (it's a browser extension installed by the user), so this skill does **simulated injection**: it navigates to the target site and evaluates the userscript body via `playwright_evaluate` / `javascript_tool` to confirm logic works against real DOM.

SKILL.md

Similar Skills

using-superpowers

180.9k

Mandates invoking relevant skills via tools before any response in coding sessions. Covers access, priorities, and adaptations for Claude Code, Copilot CLI, Gemini CLI.

3 files

superpowers

Stats

Stars0

Forks0

Last CommitMay 2, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

Test Userscript

Validate userscript behavior live in Chrome. Tampermonkey itself can't be installed/configured by Claude in Chrome (it's a browser extension installed by the user), so this skill does simulated injection: it navigates to the target site and evaluates the userscript body via playwright_evaluate / javascript_tool to confirm logic works against real DOM.

When to use

After writing or modifying a userscript
Before bumping its version with bump-version
To debug why a userscript isn't behaving on a live page

Prerequisites

The Claude in Chrome MCP must be available. Tools used:

mcp__claude-in-chrome__tabs_context_mcp — pick a tab
mcp__claude-in-chrome__tabs_create_mcp — open the target URL
mcp__claude-in-chrome__navigate — go to a @match URL
mcp__claude-in-chrome__javascript_tool — inject and run the script body
mcp__claude-in-chrome__read_console_messages — collect console.log output
mcp__claude-in-chrome__read_page / get_page_text — inspect the resulting DOM

Load each tool via ToolSearch first.

Steps

Read the target .user.js file. Parse the metadata block — extract @match / @include URLs and @grant list.
Bail early if the script needs GM_* APIs the simulator can't provide (GM_xmlhttpRequest, GM_setValue, GM_getValue, etc.). Tell the user this script can only be validated by installing it in Tampermonkey and exercising it manually. Stop here.
If @grant none, proceed: open one of the matched URLs in a new tab, wait for document-idle, then playwright_evaluate the IIFE body.
Read console messages with a pattern matching the script's name prefix to filter noise.
Read the page text or query specific selectors to confirm the expected DOM mutation occurred.
Report:
- PASS / FAIL with a one-line verdict
- What was injected, what changed, what the console said
- Any errors or warnings surfaced
If the user wants visual proof, record a short GIF of the before/after with gif_creator.

Output

A short report (under 30 lines) the user can paste into a commit message or PR description.