Selenium-Injector

• Change proxy while running (auth supported)
• remotely contoll Chrome using websockets and extensions

Feel free to test my code!

Getting Started

Dependencies

• Python >= 3.7
• Chrome-Browser installed
• Selenium

Installing

• [Windows] Install Chrome-Browser
• pip install selenium_injector

Example scripts

set proxy dynamically

from selenium_injector.webdriver import Chrome
driver = Chrome()

driver.injector.proxy.set_single(host="example_host.com", port=143, password="password", username="user-1")

driver.get("https://whatismyipaddress.com/")

driver.injector.proxy.clear()
driver.quit()

Don't forget to execute driver.quit() in the End. Else-wise your temporary folder will get flooded! and it keeps running

use events

from selenium_injector.webdriver import Chrome
import json

driver = Chrome()

driver.get("chrome://version")

t = driver.injector.socket.js.types

event_id = driver.injector.socket.make_event_id()
user = driver.injector.any_user

driver.injector.socket.exec(t.list([
    t.set_event_id(event_id),
    t.exec(
        t.path("chrome.webRequest.onCompleted.addListener"),
        args=[t.event_callback(), t.value({"urls": ["<all_urls>"]})]
    )
]), user=user, max_depth=1)

event = driver.injector.socket.event(event_id, user=user)
for e in event:  # will block forever
    e = json.loads(e)
    data = e["result"][0]
    time = e["t"]
    print(time + "\n", data['url'])

warning: as driver.quit() isn't called in this example, it will leave files in your temp directories

modify network requests

note: this is only experimental yet (not included in pypi package)

example script

from selenium_injector.webdriver import Chrome

driver = Chrome()

# modify headers
driver.injector.declarativeNetRequest.update_headers({"test": "test_2", "sec-ch-ua-platform": "Android"})
rules = driver.injector.declarativeNetRequest.dynamic_rules
headers = driver.injector.declarativeNetRequest._headers

driver.get("https://httpbin.org/headers")
input("press ENTER to continue")

# block images
driver.injector.declarativeNetRequest.update_block_on(resource_types=["image"])

driver.get("https://www.wikimedia.org/")

input("press ENTER to exit")
driver.quit()

use chrome-developer-protocoll

note: this is only experimental yet (not included in pypi package)

example script

import json
from selenium_injector.webdriver import Chrome

driver = Chrome()

dbg = driver.injector.debugger
dbg.attach()
dbg.execute("Console.enable")

events = dbg.on_event()

driver.execute_script("console.log('Hello World!')")

for event in events:
    event = json.loads(event)
    result = event["result"]
    time = event["t"]
    if result[1] == 'Console.messageAdded':
        message_text = result[2]["message"]["text"]
        print(time, message_text)
        break

driver.quit()

execute script within tab

note: this is only experimental yet (not included in pypi package)

from string

from selenium_injector.webdriver import Chrome
driver = Chrome(injector_options={"mv2":True, "mv3":True})

driver.get("https://www.wikipedia.org/")

# result only returned with mv2 extension enabled
results = driver.injector.tabs.eval_str(
            '''
            console.log(window);
            navigator.userAgent
            ''', tab_id=driver.injector.tabs.active_tab["id"])
print(results[0])

driver.quit()

with types, always returns resolves promises with MV3 automatically

from selenium_injector.webdriver import Chrome
driver = Chrome(injector_options={"mv3":True})

driver.get("https://www.wikipedia.org/")

t = driver.injector.socket.js.types
results = driver.injector.tabs.exec(t.exec(t.path("fetch"), args=[t.value("https://www.wikipedia.org/")]), timeout=40)
print(results["result"][0])

driver.quit()

find element

note: this is only experimental yet (not included in pypi package)

from selenium_injector.webdriver import Chrome
from selenium_injector.types.by import By

driver = Chrome(injector_options={"mv2":True, "mv3":True})
driver.get("https://www.wikipedia.org/")


elem = driver.injector.find_elements(By.XPATH, '//*[@id="js-link-box-en"]')
elem[0].click()

driver.quit()

Help

Please feel free to open an issue or fork!

Todo

• eval within tab scope from extension
  • mv2
    • return
    • stringify obj
  • mv3
    • return
    • stringify obj
• add events
  • make protocoll use UUIDS's
  • allow response to event within scope
    • using (...args) => {new event_handler(...args)}
• types.eval
  • for-loops
• authentificaten proxies
  • manage webrtc-leak
  • manage location api leak
  • proxy per request
• add automation tools
  • click
  • send_keys
  • find_element
    • by XPATH
• undetectability
  • make tab scripts private
  • support base_driver argument
  • make /files/js/utils.js private

Deprecated

Authors

Aurin Aegerter

License

Shield:

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Disclaimer

I am not responsible what you use the code for!!! Also no warranty!

Acknowledgments

Inspiration, code snippets, etc.

• Selenium-Profiles
• Chrome-devtools-protocol
• cdp_event_listeners
• sync websocket server
• chrome-extension-docs
• PEG-parser
• make-SV-stayalive
• stringify-obj
• inject code from mv3