Advice Request AdGuard tweaking (Yes, I confess, I did it again, changed adblocking strategy)

[Lenny_Fox]

Yes, I admit it is an addiction trying to get the most out of the combo DNS - Browser - Extensions, but to look at the bright side of life and put this in perspective, this is a decent addiction, does not cost me any money and does not offend people nor law.

I tend to switch between uBlockOrigin and Adguard extension and vica versa. I was happily using the combo NextDNS-uBlockOrigin until @oldschool and @silversurfer started to post about ClearURL and NeatURL. This triggered me to have a look at AdGuard Extension again. Benefit of Adguard is that it also has an option to check for malware (I have Smartscreen disabled because it sends full URL) PLUS AdGuard extension has a stealth mode (which is disabled by default). This stealth mode also has an URL cleaning option. The default set of URL parameters can be easily tweaked, by simply adding comma seperated parameters to the text input box.


Cleaning URL's
I added some tracking parameters of popular tracking networks in AdGuard's stealth option:

utm_source,utm_medium,utm_term,utm_campaign,utm_content,utm_name,utm_cid,utm_reader,utm_viz_id,utm_pubreferrer,utm_swu,utm_referrer,utm_social,utm_social-type,utm_place,utm_userid,utm_channel,gs_l,gs_Lcp,client,sclient,ei,sei,gws_rd,gs_gbg,gs_mss,gs_rn,ved,uact,fbclid,_hsenc,_hsmi,__hssc,__hstc,hsCtaTracking,sc_campaign,sc_channel,sc_content,sc_medium,sc_outcome,sc_geo,sc_country,elqTrackId,elqTrack,assetType,assetId,recipientId,campaignId,siteId,s_cid,itm_source,itm_medium,itm_campaign,itm_content,itm_term,spReportId,spJobID,spUserID,spMailingID,hmb_campaign,hmb_medium,hmb_source,mkt_tok,mc_cid,mc_eid,pk_campaign,pk_kwd,pk_source,pk_medium,pk_content

EDIT: it is easier to add the parameters Security123 made out of this list above in the next post!

__________

 

[HarborFront]

Yes think is in My filters add.

Hi

I added to Adguard for desktop the followings. However, in the screenshot below what's added are in green color which are different from what I added those abused TLDs previously which are in white color. Anything wrong here? And I add back ei in "URL tracking parameters to remove"?

$removeparam=ei
@@||youtube.com^$removeparam=ei
@@||googlevideo.com^$removeparam=ei

 

[petok]

Here I have also is green

I add ved now for test if work

 

[petok]

When is add one by one parameter with $removeparam= then block tracking parameter is ok, but if all parameters in one line is not working maybe I wrong
Also add this in user filters

$removeparam=gs_lcp
$removeparam=ga_source
$removeparam=uact
$removeparam=sclient

link for now is this clean
https://www.google.com/search?hl=en&gl=us&q=

 

[HarborFront]

When is add one by one parameter with $removeparam= then block tracking parameter is ok, but if all parameters in one line is not working maybe I wrong
Also add this in user filters

$removeparam=gs_lcp
$removeparam=ga_source
$removeparam=uact
$removeparam=sclient

link for now is this clean
https://www.google.com/search?hl=en&gl=us&q=

When I click your link I got this

httpx://www.google.com/webhp?gl=us

Google

Search the world's information, including webpages, images, videos and more. Google has many special features to help you find exactly what you're looking for.

www.google.com

But when I go to www.google.com it only display, without adding your

$removeparam=gs_lcp
$removeparam=ga_source
$removeparam=uact
$removeparam=sclien

httpx://www.google.com/

Halloween 2024

Happy Halloween! #GoogleDoodle

www.google.com

Mine is cleaner

 

[petok]

I use long time this link from google set to English, no locale language
The hl parameter specifies the interface language (host language) of your user interface
The gl parameter value is a two-letter country code.
The gws_rd is google web services redirect and add to ssl encryption

httpx://www.google.com/?hl=en&gl=us&gws_rd=ssl

 

[Tiamati]

According to my post #27 above fb is taken from Trace extension. You got comment on fb?

I didn't notice the FB. I thought it was wrong. Any particular reason to avoid using the other 2 i mentioned?

fb_action_ids
fb_action_types

Isn't available at Edge store so you need allowing Google ones which opens another security hole.

In that case i would pick Emsisoft Browser Security (for security, performance and privacy).

Or any of these>
Netcraft Extension
Norton Safe Web
Avira Safe Shopping

 

[HarborFront]

I didn't notice the FB. I thought it was wrong. Any particular reason to avoid using the other 2 i mentioned?

fb_action_ids
fb_action_types



In that case i would pick Emsisoft Browser Security (for security, performance and privacy).

Or any of these>
Netcraft Extension
Norton Safe Web
Avira Safe Shopping

I added

fb_action_ids
fb_action_types

to my list

 

[Lenny_Fox]

Since 1.32 uBlockOrigin also has the ability to remove URL-parameters. GorHill is planning to roll it in one of the upcoming releases.

When you want to try it out just enable "FilterAutherMode"

Spoiler: Enable URL-parameter stripping

EXAMPLE: stripping google search parameters with new QUERYPRUNE option (link uBO wiki)
||google.*^$queryprune=|ei|gs_lcp|psi|psy|uact|ved,domain=google.*

RESULT a clean search URL passing little to no info when you click on a link

PS: YEs I confess, I did it again hopping from AdGuard to uBlockOrigin this time

 

[HarborFront]

Since 1.32 uBlockOrigin also has the ability to remove URL-parameters. GorHill is planning to roll it in one of the upcoming releases.

When you want to try it out just enable "FilterAutherMode"

Spoiler: Enable URL-parameter stripping

View attachment 252959

EXAMPLE: stripping google search parameters with new QUERYPRUNE option (link uBO wiki)
||google.*^$queryprune=|ei|gs_lcp|psi|psy|uact|ved,domain=google.*

RESULT a clean search URL passing little to no info when you click on a link
View attachment 252960

PS: YEs I confess, I did it again hopping from AdGuard to uBlockOrigin this time

Does it come with a set of pre-blocked URL parameters to be removed just like in Neat URL? If not you'll need to craft the queryprune filters yourself

As quoted from the uBO wiki link

Poorly crafted queryprune filters can have deleterious effects on performance, experienced filter authors are expected to understand well how to craft optimal filters.

 

[forgottenuser79643]

Since 1.32 uBlockOrigin also has the ability to remove URL-parameters. GorHill is planning to roll it in one of the upcoming releases.

When you want to try it out just enable "FilterAutherMode"

Spoiler: Enable URL-parameter stripping

View attachment 252959

EXAMPLE: stripping google search parameters with new QUERYPRUNE option (link uBO wiki)
||google.*^$queryprune=|ei|gs_lcp|psi|psy|uact|ved,domain=google.*

RESULT a clean search URL passing little to no info when you click on a link
View attachment 252960

PS: YEs I confess, I did it again hopping from AdGuard to uBlockOrigin this time

In regards to the scriptlets question, i do not know as i never used that line of code. BUT if it's just for AdGuard maintained filterlists. You can find the uBlock optimized filterlists maintained by the AdGuard team themselves.

Example: https://filters.adtidy.org/extension/ublock/filters/3.txt replace the ublock part in the url with chromium for the AdGuard compatible list or the other way around for the uBlock compatible list.

 

[petok]

Here other variant for cleaning url in google no destroy other youtube video for AdGuard for Windows
I test this 2 days and now post here (if before have rules from my post up remove )

||google.*^$removeparam=ei|gs_lcp|oq|sclient|source|uact|ved

 

[Lenny_Fox]

Does it come with a set of pre-blocked URL parameters to be removed just like in Neat URL? If not you'll need to craft the queryprune filters yourself

As quoted from the uBO wiki link

Poorly crafted queryprune filters can have deleterious effects on performance, experienced filter authors are expected to understand well how to craft optimal filters.

1. No as far as I know you have to construct them yourself. As with many other advanced uBO features, in time they will appear in the uBO build-in lists.

2. Good point better to change the rule for AG compatibility (thx @petok uBO also recognises AG's removeparam]) and reduced performance overhead:

||google.*^$removeparam=ei|gs_lcp|oq|sclient|source|uact|ved|psi|psy,domain=google.*

 

[petok]

I update just add others parameter and others just need test 4 new parameters to see what is loading then I will add to rule.
I see google have tricks in loading pages for now this is update thanks for other member for parameter

||google.*^$removeparam=aq|ei|gs_l|gs_lcp|oq|pq|psi|psy|sa|sclient|source|sourceid|uact|ved

@Lenny_Fox why is add domain=google.* on end ?

 

[Lenny_Fox]

That is for uBlockOrigin. It applies the rule only on domains when visiting a website with google as domain name for any country code (* is wildcard) e.g. on google.com, google.fr, google.de etc.

This way the rule is only processed when seaching with google

 

[SeriousHoax]

Hi guys. Yuki has something to say about the discussion going on here. I recommend you to take a look.

[English only] Questions about uBlock Origin/AdGuard · Issue #4 · Yuki2718/adblock

hi there yuki, hope you're doing ok. so, i need your help. do you think it's a bad idea to globally noop "wp.com, wordpress.com, wix.com, parastorage.com" and similar commonly used cms's with frequ...

github.com

 

[Lenny_Fox]

I thank @SeriousHoax for attending me on Yuki's comments. Fun-fact when I read Yuki's earlier comments about HTTP third-party rule I did not understand what he meant, now he worded it differently i think I understand what he is advising. Same with $queryprune (now $removeparam), first time I followed the documentation at that time (see note 1), obviously something was depreciated according to current documentation (see note 2).

It would be more helpful to write a better manual than complaining about wrong usage and people being to dumb to use advanced rules features. Just giving it a second go and hopefully his Yuki-ness will comment on the rules below for educational purpose.
______________________________________________________________________
||google.nl/search?*aq$removeparam=aq,domain=google.nl
||google.nl/search?*ei$removeparam=ei,domain=google.nl
||google.nl/search?*gs_l$removeparam=gs_l,domain=google.nl
||google.nl/search?*gs_lcp$removeparam=gs_lcp,domain=google.nl
||google.nl/search?*oq$removeparam=oq,domain=google.nl
||google.nl/search?*pq$removeparam=pq,domain=google.nl
||google.nl/search?*psi$removeparam=psi,domain=google.nl
||google.nl/search?*psy$removeparam=psy,domain=google.nl
||google.nl/search?*sa$removeparam=sa,domain=google.nl
||google.nl/search?*sclient$removeparam=sclient,domain=google.nl
||google.nl/search?*source$removeparam=source,domain=google.nl
||google.nl/search?*sourceid$removeparam=sourceid,domain=google.nl
||google.nl/search?*uact$removeparam=uact,domain=google.nl
||google.nl/search?*ved$removeparam=ved,domain=google.nl
_______________________________________________________________________
In stead of writing one general purpose rule, I made it specific to google.nl, I also split the different parameters into individual rules so a rule has more tokens and they can be put in a token-bucket for more efficient processing.

Note1: I added a picture of an actual rule of uBo build-in filters (using the pipe) to show using pipe (|) in version one was according the uBO wiki-documentation at that time
Note2: I added a picture of current documentation. so everyone can check the version two rules are according the documentation available at the moment of writing this post

 

[Lenny_Fox]

Follow up post, actual load time testing of removeparam

I compared this with the multi-rule version with single rule version. The multi-rule versions were written in exactly the same manner as the queryprune example I found in the uBlock build-in list. Assuming the author of the build-in uBO filter knows how to write efficient rules, the copied multi-rule version should also be efficient.

When using uBO's build in logger I could not notice any performance differences (I repoeated the test 5 times)

Single ineffecient version removeparam

Muti-rule removeparam version following structure and syntax of queryprune rule of uBO build-in filter

So using the uBO logger ihe inefficient rule performed the same as the efficient rule. So I decided to use another tool and to give credit to Yuki the multi-rule version was more efficient (4 digit after the decimal break). So a single in efficient line will cost 0.0001 to 0.0002 response time. Which is a lot when uBo easily processes half a million rules in half a second.

Lesson learned: we need crafted rule writers in this forum like Yuki

EDIT this evening I researched what info these parameters contained, so only kept two removeparam's and dumped all others (I am not paranoid, just playing with stuff so it is time to lower my level of excitement on the new found feature and only use it when it really increases privacy). A

! strip google click identifier
||google.nl/*glcid$removeparam=glid,domain=google.nl
! strip google search location parameter
||google.nl/*gs_lcp$removeparam=gs_lcp,domain=google.nl

From a performanc epoint of view these rules worked as good (4 digits behind decimal point) or other factors influencing response time had more impact (other network traffic and/or available bandwith) to notice the difference.

 

[Lenny_Fox]

Sorry I made a typing mistake correct rules are

! strip google click identifier
||google.nl/*gclid$removeparam=gclid,domain=google.nl
! strip google search location
||google.nl/*gs_lcp$removeparam=gs_lcp,domain=google.nl

 

[HarborFront]

Sorry I made a typing mistake correct rules are

! strip google click identifier
||google.nl/*gclid$removeparam=gclid,domain=google.nl
! strip google search location
||google.nl/*gs_lcp$removeparam=gs_lcp,domain=google.nl

I'm not familiar with what you are doing. However, using one example

||google.nl/*gs_lcp$removeparam=gs_lcp,domain=google.nl

I can see that the google.nl (front and back) refers to searches in Netherlands. Am I right?

But if I want to search US, UK and other countries can use a wild card like google.* instead of a specific country? Google searches are used if you use Google as the search engine, right? If I'm using say DDG, Qwant etc then no need of them, correct?

Thanks

 

[forgottenuser79643]

I'm not familiar with what you are doing. However, using one example

||google.nl/*gs_lcp$removeparam=gs_lcp,domain=google.nl

I can see that the google.nl (front and back) refers to searches in Netherlands. Am I right?

But if I want to search US, UK and other countries can use a wild card like google.* instead of a specific country? Google searches are used if you use Google as the search engine, right? If I'm using say DDG, Qwant etc then no need of them, correct?

Thanks

Whether you use ||google.[<TLD-here>] or ||google.* doesn't matter in uBlock from what I know, but if you exclusively visit a particular TLD variant of Google than it might be more efficient for you to fill in that one, no? I personally have always with go google.* unless I was looking to either block, hide or allow something very specific when it comes to Google.