I have reviewed the following documentation in the AdGuard public Knowledge Base

Adguard Filters Knowledge Bases Review.md

Adguard Filters Knowledge Bases Review

I have reviewed the following documentation in the AdGuard public Knowledge Base: https://adguard.com/kb

• Types of filtering rules – How ad blocking works
• AdGuard filter policy | AdGuard Knowledge Base
• How to create your own ad filters | AdGuard Knowledge Base

I have not yet reviewed the AdGuard DNS KB. However, I previously attempted to contribute to the List-KR repository by blocking KakaoTalk in-app advertising domains:

annoyance: com.kakao.talk by mu-hun · Pull Request #1019 · List-KR

Self-Note: HTML Filtering rules

Among the Types of filtering rules, the "HTML filtering rules" section is said to be limited depending on the platform:

These rules are quite complicated and require the ad blocker to be granted certain access rights, so not all platforms support them. Currently, these rules work only in the AdGuard Firefox add-on and in the AdGuard apps for Windows, Mac, and Android.

I was curious about how these rules are used in practice, so I looked for an example from the AdGuard filters repositories:

(Reference example inspired by rules used in the AdGuard Filters repositories on GitHub.)

Example

Example of an HTML filtering rule used in production filters:

example.com$$script[tag-content="adblock"][max-length="20000"]

What this rule does

1. Targets pages on example.com.
2. Scans the raw HTML response before the browser parses it.
3. Removes inline <script> elements containing the text adblock when they match the specified conditions.

Why HTML filtering is needed here

In some cases, websites embed anti-adblock logic directly inside large inline scripts within the HTML document:

<script>
  if (window.adblockDetected) {
    document.body.innerHTML = 'Please disable your ad blocker'
  }
</script>

This script cannot be blocked with a normal network rule because it is not loaded from a separate URL. A cosmetic rule would also be unreliable, since the script executes immediately while the page is being parsed.

Using an HTML filtering rule removes the problematic script before it is parsed and executed by the browser, preventing the anti‑adblock logic from running.

This is why HTML filtering rules are generally treated as a last‑resort technique and used only when blocking requests or cosmetic filtering cannot solve the problem.

For this reason, HTML filtering rules are typically used only when the problem cannot be solved with network or cosmetic filtering.

Comparison: Three main rules

Three main rules: Basic (network) rules, Cosmetic rules, and HTML filtering rules.

Category	Basic rules (Network filtering)	Cosmetic rules (Element hiding)	HTML filtering rules
Primary purpose	Block requests to ads, trackers, or unwanted resources	Hide or remove visible page elements after the page loads	Remove or rewrite parts of the HTML document before it is parsed
Execution stage	Before resource download	After DOM is created	Before HTML is parsed and executed
What it operates on	Network requests (URLs)	DOM elements rendered in the page	Raw HTML response body
Typical syntax	/ads.js	##.ads	example.com#$#remove(".ads"), example.com$$script[tag-content="ad"]
Typical use cases	Blocking ad scripts, trackers, ad images, analytics	Hiding banners, sponsored blocks, cookie notices	Removing inline scripts, anti-adblock logic, injected paywalls
Performance impact	Very low (fastest method)	Low	Higher (requires HTML scanning/modification)
Platform support	All AdGuard products	All AdGuard products	Limited platforms only
Reliability	Highly reliable if the request URL is identifiable	Works well for static page elements	Useful for complex inline logic that other rules cannot stop
Preferred usage order	1st choice	2nd choice	Last resort

This table reflects how filter maintainers typically think about them when deciding which mechanism to use.

Notes from "Inside AGLint v4" (ADFS presentation)

I watched the AGLint v4 presentation at last month. It introduced Scriptlets and showed examples of the AutoFix and Suggestion features.

Scriptlet — modify or disable specific script functionality on a page

example.com#%$//scriptlet('abort-on-property-read' 'alert')

Stops a script on example.com if it attempts to access the alert browser API.

AGLint feature — removing duplicated CSS declarations

#$#.foo { padding: 1px; padding 1px; }

In this case, AGLint can automatically remove one of the duplicated declarations (AutoFix).

#$#.foo { padding: 1px; padding 1px; }

If it is unclear which declaration should be kept, AGLint provides suggestions instead of applying an automatic fix.