phrazzld

browser-extension-dev

2
1
# Install this skill:
npx skills add phrazzld/claude-config --skill "browser-extension-dev"

Install specific skill from multi-skill repository

# Description

|

# SKILL.md


name: browser-extension-dev
description: |
Browser extension development with security and cross-browser support. Use when:
- Building Chrome, Firefox, or Safari extensions
- Requesting permissions in manifest
- Implementing content scripts or background workers
- Handling cross-browser compatibility
- Planning extension updates
Keywords: browser extension, Manifest V3, content script, background script,
permissions, Chrome extension, Firefox addon, WebExtensions API


Browser Extension Development

Manifest V3, minimal permissions, cross-browser first.

Manifest V3

Required for new extensions:

{
  "manifest_version": 3,
  "name": "My Extension",
  "version": "1.0.0",
  "permissions": ["storage"],
  "host_permissions": ["https://api.example.com/*"],
  "background": {
    "service_worker": "background.js"
  },
  "content_scripts": [{
    "matches": ["https://*.example.com/*"],
    "js": ["content.js"]
  }],
  "content_security_policy": {
    "extension_pages": "script-src 'self'; object-src 'none'"
  }
}

Minimal Permissions

Request only what you need. Justify each permission:

Permission Use Case
storage Save user preferences
activeTab Access current tab on user action
scripting Inject scripts programmatically
https://specific.com/* Access specific API

Never: <all_urls> without documented justification.

Use optional permissions for non-core features:

chrome.permissions.request({
  origins: ['https://optional.com/*']
}, (granted) => {
  if (granted) enableFeature();
});

Message Passing

Always validate messages at boundaries:

// background.js
chrome.runtime.onMessage.addListener((message, sender, sendResponse) => {
  // Validate sender
  if (!sender.tab) return;

  // Validate message structure
  if (message.type !== 'FETCH_DATA') return;
  if (typeof message.url !== 'string') return;

  // Process safely
  fetchData(message.url).then(sendResponse);
  return true; // async response
});

Never trust data from content scripts without validation.

Cross-Browser Support

// Use webextension-polyfill for API normalization
import browser from 'webextension-polyfill';

// Feature detection
if (browser.storage?.sync) {
  await browser.storage.sync.set({ key: value });
} else {
  await browser.storage.local.set({ key: value });
}

Test on Chrome, Firefox, Safari, Edge.

Updates

  • Semantic versioning (MAJOR.MINOR.PATCH)
  • Changelog accessible in-extension
  • Gradual rollout for major changes
  • Highlight permission changes to users
  • Maintain backward compatibility
  • Preserve user data across versions

Anti-Patterns

  • <all_urls> permission hoarding
  • eval() or remote code execution
  • Trusting content script data without validation
  • Manifest V2 for new development
  • Chrome-only without cross-browser testing
  • Silent permission scope creep

# Supported AI Coding Agents

This skill is compatible with the SKILL.md standard and works with all major AI coding agents:

Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.