muqsitnawaz/.agents-system/browser

Claude Code and agent skill by muqsitnawaz/.agents-system

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Drive a browser to automate websites — fill forms, click buttons, take screenshots, scrape pages. Uses the built-in `browser` command (or `agents browser`).

Install

npx vskill@latest install muqsitnawaz/.agents-system/browser

SKILL.md

Browser Automation

The browser command (shorthand for agents browser) provides CDP-based browser automation with profile and task management.

Quick Start

bash
# Create a profile (one-time setup)
browser profiles create my-profile -b chrome -e cdp://localhost:9222
 
# Start a task (auto-launches browser if needed)
browser start my-task --profile my-profile
 
# Navigate and interact
browser navigate my-task https://example.com
browser refs my-task              # Get clickable element refs
browser click my-task <tabId> <ref>
browser type my-task <tabId> <ref> "hello"
browser screenshot my-task
browser stop my-task

Profiles

Profiles define browser type and connection endpoint. Create once, reuse across tasks.

bash
# Local Chrome
browser profiles create local -b chrome -e cdp://localhost:9222
 
# Remote browser via SSH tunnel
browser profiles create mac-mini -b comet -e ssh://mac-mini?port=9333
 
# List profiles
browser profiles list

Supported browsers: chrome, comet, chromium, brave, edge

Tasks

Tasks are browser sessions. Each task gets its own tabs and state.

bash
browser start [task-name] --profile <profile>   # auto-generates name if omitted
browser tasks                                    # list active tasks
browser stop <task>                              # close task and its tabs

Navigation & Tabs

bash
browser navigate <task> <url>          # open URL in new tab
browser tabs <task>                    # list open tabs
browser close <task> [tabId]           # close specific tab or all

DOM Interaction

bash
browser refs <task> [tabId]                    # get interactive element refs
browser click <task> <tabId> <ref>             # click element
browser type <task> <tabId> <ref> "text"       # type into element
browser press <task> <tabId> Enter             # press key (Enter, Tab, Escape)
browser hover <task> <tabId> <ref>             # hover over element

Screenshots & Evaluation

bash
browser screenshot <task> [tabId]              # capture to ~/.agents-system/browser/sessions/<task>/
browser evaluate <task> <tabId> "document.title"   # run JS, return result

Account Credentials

Before navigating to sites that require login, load credentials:

bash
eval "$(agents secrets export browser-accounts --plaintext)"

Known accounts (configure in browser-accounts bundle):

ServiceUsername variablePassword variable
Grafana$GRAFANA_USERNAME$GRAFANA_PASSWORD
Cloudflare$CLOUDFLARE_USERNAME$CLOUDFLARE_PASSWORD

Always screenshot first — if the session is still alive in the profile, skip login.

Workflow Pattern

  1. Start a task with a profile
  2. Navigate to target URL
  3. Refs to see clickable elements with their ref numbers
  4. Click / type / press using refs
  5. Refs again after each action — refs are ephemeral
  6. Screenshot liberally — screenshots are your eyes
  7. Stop the task when done

Rich Text Editors

type may fail on contenteditable / ProseMirror. Use evaluate:

bash
browser evaluate <task> <tabId> 'document.execCommand("insertText", false, "your text")'

Remote Browsers (SSH)

For browsers on remote machines:

bash
# Create profile with SSH endpoint
browser profiles create remote-mac -b comet -e ssh://user@hostname?port=9222
 
# Use normally — SSH tunnel is automatic
browser start task --profile remote-mac
browser navigate task https://example.com

The SSH driver launches the browser on the remote host and tunnels CDP back.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

See the full security report

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

See evaluation results

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/muqsitnawaz/.agents-system/browser/badge)](https://verified-skill.com/skills/muqsitnawaz/.agents-system/browser)

Something wrong with this skill or its page? Tell us.