# QwenPaw Browser

Drives a live browser with async Python through QwenPaw's built-in SDK.

- Repository: https://github.com/agentscope-ai/QwenPaw (37135 stars)
- Category: Engineering
- License: Apache-2.0
- Last updated: 2026-10-30
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add agentscope-ai/QwenPaw --skill browser -g -a claude-code -y`
Codex: `npx skills add agentscope-ai/QwenPaw --skill browser -g -a codex -y`
Cursor: `npx skills add agentscope-ai/QwenPaw --skill browser -g -a cursor -y`
Gemini CLI: `npx skills add agentscope-ai/QwenPaw --skill browser -g -a gemini-cli -y`

## Description

Drive a live browser with async Python against QwenPaw's builtin Browser SDK. The full reference is below; re-load this browser skill after context compaction.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# Browser

Work with discipline: perceive the current page, act through the documented
surface, then re-perceive before claiming success. State only facts observed
in this turn. When stuck, a complete delivery is the step you are stuck on
plus the partial results you have verified, never fill in content you did not
observe just to produce a full answer.

Respect the human boundary: for login, captcha, 2FA, or any human-only step,
call `await browser.handoff(...)` and stop. Never automate those flows.
When the browser cannot finish, do not substitute data from another channel
(such as web_search) and still call it a browser result, state each fact's
real source.

This is QwenPaw's builtin Browser SDK, not Playwright. Its surface is closed:
anything not listed does not exist. The complete reference is below; re-load
this browser skill with the Skill tool if it is no longer in context.

<!-- BEGIN GENERATED: browser-manual -->
QwenPaw Browser SDK, complete reference. This is QwenPaw's OWN internal
SDK and this is the ENTIRE API; these are all the entrypoints. The SDK
is already in scope as Browser, call the methods below directly. Write
async Python. Work in a loop: perceive → act → verify.

The rest: https://github.com/agentscope-ai/QwenPaw/blob/HEAD/src/qwenpaw/agents/skills/browser-en/SKILL.md