Splash authentication credentials potentially leaked to target websites

Impact

If you use HttpAuthMiddleware (i.e. the http_user and http_pass spider attributes) for Splash authentication, any non-Splash request will expose your credentials to the request target. This includes robots.txt requests sent by Scrapy when the ROBOTSTXT_OBEY setting is set to True.

Patches

Upgrade to scrapy-splash 0.8.0 and use the new SPLASH_USER and SPLASH_PASS settings instead to set your Splash authentication credentials safely.

Workarounds

If you cannot upgrade, set your Splash request credentials on a per-request basis, using the splash_headers request parameter, instead of defining them globally using the HttpAuthMiddleware.

Alternatively, make sure all your requests go through Splash. That includes disabling the robots.txt middleware.

For more information

If you have any questions or comments about this advisory:

• Open an issue
• Email us