feat: Use NavigatorUAData and navigator.webdriver to improve bot detection

cypress/e2e/ua.cy.ts

@@ -0,0 +1,14 @@
+/// <reference types="cypress" />
+import { start } from '../support/setup'
+
+describe('User Agent Blocking', () => {
+    it('should pick up that our automated cypress tests are indeed bot traffic', async () => {
+        cy.skipOn('windows')
+        start({})
+
+        cy.window().then((win) => {
+            const isLikelyBot = win.eval('window.posthog._is_bot()')
+            expect(isLikelyBot).to.eql(true)
+        })
+    })
+})

cypress/support/commands.ts

@@ -13,6 +13,7 @@ Cypress.Commands.add('posthogInit', (options) => {
             $captures.push(event)
             $fullCaptures.push(eventData)
         },
+        opt_out_useragent_filter: true,
         ...options,
     })
 })

cypress/support/e2e.ts

@@ -1,4 +1,5 @@
 import './commands'
+import '@cypress/skip-test/support'
 
 // Add console errors into cypress logs.
 Cypress.on('window:before:load', (win) => {

cypress/support/index.ts

@@ -2,6 +2,7 @@
 
 import { PostHog } from '../../src/posthog-core'
 import { PostHogConfig } from '../../src/types'
+import '@cypress/skip-test'
 
 declare global {
     // eslint-disable-next-line @typescript-eslint/no-namespace

cypress/support/setup.ts

@@ -39,7 +39,10 @@ export const start = ({
     cy.visit(url)
 
     if (initPosthog) {
-        cy.posthogInit(options)
+        cy.posthogInit({
+            opt_out_useragent_filter: true, // we ARE a bot, so we need to enable this opt-out
+            ...options,
+        })
     }
 
     if (resetOnInit) {

cypress/tsconfig.json

@@ -1,7 +1,7 @@
 {
     "compilerOptions": {
         "target": "es2015",
-        "lib": ["es5", "dom"],
+        "lib": ["es5", "dom", "es2015"],
         "types": ["cypress", "node"],
         "moduleResolution": "node"
     },

package.json

@@ -43,6 +43,7 @@
         "@babel/plugin-transform-react-jsx": "^7.23.4",
         "@babel/preset-env": "7.18.9",
         "@babel/preset-typescript": "^7.18.6",
+        "@cypress/skip-test": "^2.6.1",
         "@jest/globals": "^27.5.1",
         "@rollup/plugin-babel": "^6.0.4",
         "@rollup/plugin-json": "^6.1.0",
@@ -64,8 +65,8 @@
         "babel-eslint": "10.1.0",
         "babel-jest": "^26.6.3",
         "compare-versions": "^6.1.0",
-        "cypress": "13.6.3",
-        "cypress-localstorage-commands": "^2.2.5",
+        "cypress": "13.13.2",
+        "cypress-localstorage-commands": "^2.2.6",
         "date-fns": "^3.6.0",
         "eslint": "8.56.0",
         "eslint-config-posthog-js": "link:eslint-rules",

playground/nextjs/package.json

@@ -4,7 +4,7 @@
     "private": true,
     "scripts": {
         "clean-react": "cd ../../react && rm -rf ./node_modules/",
-        "dev": "pnpm run link-posthog-js && pnpm run clean-react && next dev --experimental-https",
+        "dev": "pnpm run link-posthog-js && pnpm run clean-react && next dev",
         "dev-crossdomain": "pnpm run link-posthog-js && pnpm run clean-react && NEXT_PUBLIC_CROSSDOMAIN=1 next dev --experimental-https",
         "build": "pnpm run build-posthog-js && pnpm run link-posthog-js && pnpm run clean-react && next build",
         "start": "next start",

playground/nextjs/pages/ua.tsx

@@ -0,0 +1,32 @@
+import { useEffect, useState } from 'react'
+
+// Try this page with some of the following commands:
+// chrome --headless --disable-gpu --print-to-pdf http://localhost:3000/ua --virtual-time-budget=10000
+// chrome --headless --disable-gpu --print-to-pdf http://localhost:3000/ua --virtual-time-budget=10000 --user-agent="RealHuman"
+
+export default function Home() {
+    const [isClient, setIsClient] = useState(false)
+    useEffect(() => {
+        setIsClient(true)
+    }, [])
+    if (!isClient) {
+        return <pre>Not client</pre>
+    }
+    return (
+        <dl>
+            <dt>UA</dt>
+            <dd>
+                <code>{navigator.userAgent}</code>
+            </dd>
+            <dt>WebDriver</dt>
+            <dd>
+                <code>{String(navigator.webdriver)}</code>
+            </dd>
+            <dt>NavigatorUAData brands</dt>
+            <dd>
+                {/* eslint-disable-next-line compat/compat */}
+                <code>{JSON.stringify((navigator as any).userAgentData?.brands)}</code>
+            </dd>
+        </dl>
+    )
+}

src/__tests__/posthog-core.ts

@@ -144,24 +144,28 @@ describe('posthog core', () => {
         })
 
         it('respects opt_out_useragent_filter (default: false)', () => {
-            const originalUseragent = globals.userAgent
-            ;(globals as any)['userAgent'] =
-                'Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Googlebot/2.1; +http://www.google.com/bot.html) Chrome/W.X.Y.Z Safari/537.36'
-
+            const originalNavigator = globals.navigator
+            ;(globals as any).navigator = {
+                ...globals.navigator,
+                userAgent:
+                    'Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Googlebot/2.1; +http://www.google.com/bot.html) Chrome/W.X.Y.Z Safari/537.36',
+            }
             const hook = jest.fn()
             const posthog = posthogWith(defaultConfig, defaultOverrides)
             posthog._addCaptureHook(hook)
 
             posthog.capture(eventName, {}, {})
             expect(hook).not.toHaveBeenCalledWith('$event')
-            ;(globals as any)['userAgent'] = originalUseragent
+            ;(globals as any)['navigator'] = originalNavigator
         })
 
         it('respects opt_out_useragent_filter', () => {
-            const originalUseragent = globals.userAgent
-
-            ;(globals as any)['userAgent'] =
-                'Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Googlebot/2.1; +http://www.google.com/bot.html) Chrome/W.X.Y.Z Safari/537.36'
+            const originalNavigator = globals.navigator
+            ;(globals as any).navigator = {
+                ...globals.navigator,
+                userAgent:
+                    'Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Googlebot/2.1; +http://www.google.com/bot.html) Chrome/W.X.Y.Z Safari/537.36',
+            }
 
             const hook = jest.fn()
             const posthog = posthogWith(
@@ -184,7 +188,7 @@ describe('posthog core', () => {
                 })
             )
             expect(event.properties['$browser_type']).toEqual('bot')
-            ;(globals as any)['userAgent'] = originalUseragent
+            ;(globals as any)['navigator'] = originalNavigator
         })
 
         it('truncates long properties', () => {

src/__tests__/utils.test.ts

@@ -9,7 +9,8 @@
 
 import { _copyAndTruncateStrings, isCrossDomainCookie, _base64Encode } from '../utils'
 import { Info } from '../utils/event-utils'
-import { isBlockedUA, DEFAULT_BLOCKED_UA_STRS } from '../utils/blocked-uas'
+import { isLikelyBot, DEFAULT_BLOCKED_UA_STRS, isBlockedUA, NavigatorUAData } from '../utils/blocked-uas'
+import { expect } from '@jest/globals'
 
 function userAgentFor(botString: string) {
     const randOne = (Math.random() + 1).toString(36).substring(7)
@@ -103,13 +104,13 @@ describe('utils', () => {
         })
     })
 
-    describe('user agent blocking', () => {
+    describe('isLikelyBot', () => {
         it.each(DEFAULT_BLOCKED_UA_STRS.concat('testington'))(
             'blocks a bot based on the user agent %s',
             (botString) => {
                 const randomisedUserAgent = userAgentFor(botString)
 
-                expect(isBlockedUA(randomisedUserAgent, ['testington'])).toBe(true)
+                expect(isLikelyBot({ userAgent: randomisedUserAgent } as Navigator, ['testington'])).toBe(true)
             }
         )
 
@@ -125,10 +126,93 @@ describe('utils', () => {
             [
                 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/125.0.6422.175 Safari/537.36 (compatible; Google-HotelAdsVerifier/2.0)',
             ],
+            [
+                'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) HeadlessChrome/122.0.0.0 Safari/537.36',
+            ],
+            [
+                'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Cypress/13.6.3 Chrome/114.0.5735.289 Electron/25.8.4 Safari/537.36',
+            ],
         ])('blocks based on user agent', (botString) => {
             expect(isBlockedUA(botString, [])).toBe(true)
             expect(isBlockedUA(botString.toLowerCase(), [])).toBe(true)
             expect(isBlockedUA(botString.toUpperCase(), [])).toBe(true)
+            expect(isLikelyBot({ userAgent: botString } as Navigator, [])).toBe(true)
+            expect(isLikelyBot({ userAgent: botString.toLowerCase() } as Navigator, [])).toBe(true)
+            expect(isLikelyBot({ userAgent: botString.toUpperCase() } as Navigator, [])).toBe(true)
+        })
+
+        it.each([
+            ['Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:129.0) Gecko/20100101 Firefox/129.0'],
+            [
+                'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.0.0 Safari/537.36',
+            ],
+            [
+                'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.5 Safari/605.1.15',
+            ],
+            [
+                'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) elec/1.0.0 Chrome/126.0.6478.127 Electron/31.2.1 Safari/537.36',
+            ],
+        ])('does not block based on non-bot user agent', (userAgent) => {
+            expect(isBlockedUA(userAgent, [])).toBe(false)
+            expect(isBlockedUA(userAgent.toLowerCase(), [])).toBe(false)
+            expect(isBlockedUA(userAgent.toUpperCase(), [])).toBe(false)
+            expect(isLikelyBot({ userAgent } as Navigator, [])).toBe(false)
+            expect(isLikelyBot({ userAgent: userAgent.toLowerCase() } as Navigator, [])).toBe(false)
+            expect(isLikelyBot({ userAgent: userAgent.toUpperCase() } as Navigator, [])).toBe(false)
+        })
+
+        it('blocks based on the webdriver property being set to true', () => {
+            expect(isLikelyBot({ webdriver: true } as Navigator, [])).toBe(true)
+        })
+
+        it('blocks based on userAgentData', () => {
+            const headlessUserAgentData: NavigatorUAData = {
+                brands: [
+                    { brand: 'Not)A;Brand', version: '99' },
+                    { brand: 'HeadlessChrome', version: '127' },
+                    { brand: 'Chromium', version: '127' },
+                ],
+            }
+            expect(
+                isLikelyBot(
+                    {
+                        userAgentData: headlessUserAgentData,
+                    } as Navigator,
+                    []
+                )
+            ).toBe(true)
+        })
+
+        it('does not block a normal browser based of userAgentData', () => {
+            const realUserAgentData: NavigatorUAData = {
+                brands: [
+                    { brand: 'Not)A;Brand', version: '99' },
+                    { brand: 'Google Chrome', version: '127' },
+                    { brand: 'Chromium', version: '127' },
+                ],
+            }
+            expect(
+                isLikelyBot(
+                    {
+                        userAgentData: realUserAgentData,
+                    } as Navigator,
+                    []
+                )
+            ).toBe(false)
+        })
+
+        it('does not crash if the type of navigatorUAData changes', () => {
+            // we're not checking the return values of these, only that they don't crash
+            // @ts-expect-error testing invalid data
+            isLikelyBot({ userAgentData: { brands: ['HeadlessChrome'] } } as Navigator, [])
+            // @ts-expect-error testing invalid data
+            isLikelyBot({ userAgentData: { brands: [() => 'HeadlessChrome'] } } as Navigator, [])
+            isLikelyBot({ userAgentData: { brands: () => ['HeadlessChrome'] } } as unknown as Navigator, [])
+            isLikelyBot({ userAgentData: 'HeadlessChrome' } as unknown as Navigator, [])
+            isLikelyBot({ userAgentData: {} } as unknown as Navigator, [])
+            isLikelyBot({ userAgentData: null } as unknown as Navigator, [])
+            isLikelyBot({ userAgentData: () => ['HeadlessChrome'] } as unknown as Navigator, [])
+            isLikelyBot({ userAgentData: true } as unknown as Navigator, [])
         })
     })