How to avoid Puppeteer being detected by front-end JS

Tools and information

• QQ Group-Javascript advanced crawler https://jq.qq.com/?_wv=1027&k=5Bcu3YU-The author's self-built group, welcome to join!
• China trademark network encryption interface https://github.com/rockswang/wsjs.saic.mmewmd-for   demonstration only
• awesome-jave-crawler https://github.com/rockswang/awesome-java-crawler-crawler  related tools and materials collected by the author

Preface

I have been watching puppeteer in the past two days and found that it can also be detected by a certain number of front-end js!?

I searched the issue area of ​​github. It turns out that there is navigator.webdriver attribute in the chrome started by puppeteer. What's the matter? The foreigner really needs to set up the torii...

plan 1

The solutions seen in the issue area:

await this.page.evaluateOnNewDocument(() => {
        Object.defineProperty(navigator, 'webdriver', {
        get: () => undefined,
    });
}

But to be honest, this is still a bit problematic, because it can still be detected with "webdriver" in navigator.

Scenario 2

I want to find out which link to add the "webdriver" attribute, but text search found that there is no puppeteer source code...

Later, I looked at the command line startup parameters with chrome:version in chrome, and found this: "--enable-automation" ......

There is a problem at first glance. In order to confirm, I searched for the description of this command line parameter:

--enable-automation: Inform users that their browser is being controlled by an automated test.

After confirming that it can be killed, add an ignore default parameter when starting chrome:

const browser = await puppeteer.launch({ignoreDefaultArgs: ["--enable-automation"]});