AdGuard Filter Auto-Generation Tool

Overview

This repository provides a set of scripts to:

• Generate block lists for AdGuard DNS.
• Curate domain lists considered harmful to children in Japanese-speaking regions.
• Create separate files for each category and merge them into a single consolidated list.
• Extract and list only those domains not already included in the public block lists published via filters.json.

The scripts download source block lists, compare them against target files, and extract newly appearing domains for use in AdGuard DNS.

Requirements

• bash
• wget
• md5sum (coreutils)
• Python 3.6 or higher
• Python standard libraries: argparse, glob, datetime

Installation

git clone https://github.com/<your-username>/AdGuardFilters.git
cd AdGuardFilters
chmod +x make_all_filter.sh BlockLists/download_all.sh

Scripts

make_all_filter.sh

1. Change into the BlockLists directory and remove the existing downloads folder.
2. Run ./download_all.sh to fetch all BlockLists.
3. Return to the root directory and execute filter_domains.py on the specified target files.

./make_all_filter.sh

BlockLists/download_all.sh

• Fetches the filters.json file from the AdGuard HostlistsRegistry and extracts each downloadUrl entry.
• Dependencies: bash, wget, jq, coreutils (provides gmd5sum or md5).
• Downloads all extracted URLs into BlockLists/downloads/, naming each file with its MD5 hash and original basename.
• Logs any failed download URLs to BlockLists/download_failed.log.

filter_domains.py

• Reads all existing domain lists in BlockLists/downloads/.
• Compares them against one or more target files provided as arguments.
• Writes non-duplicate domains for each target to output/<target>_filters.txt.
• Checks the validity of each new domain candidate by querying the AdGuard DNS family resolver API (https://family.adguard-dns.com/resolve?name=<domain>&type=A). Only domains with Status=0 are retained.
• Aggregates all valid new domains into output/all_filters.txt, prefixed by a header comment with metadata.

Output Files

• output/<target>_filters.txt: New domains for each individual target file.
• output/all_filters.txt: Consolidated list of all new domains across targets, with a header containing title, description, homepage, and last modified timestamp.

Usage

1. Execute make_all_filter.sh to perform the full download-and-filter pipeline in one command.
2. Inspect the generated files in the output/ directory.

License

This project is released under the MIT License.