Automating Ad commits
summary
A puppeteer-based network scanner (NWSS.js) automatically crawls target sites, applies JSON “recipes” of regex + string + whois/dig checks, and emits adblock and host-format rules; deployed as a local runner or GitHub Action it produces screenshots, headful/headless runs, and DNS/PI-hole outputs. Over four months it yields an 8–10x increase in additions, routinely adds ~300 domains per run while unattended, and reduces human maintenance at the cost of careful private recipe tuning.
AI-written summary · about
chapters (20)
- 0:00Introduction and welcome
- 1:02Problem with ad server randomization
- 7:30Previous methods and their downsides
- 10:22Introducing the network scanning script
- 11:26Benefits of the automated script
- 13:55How the script works with examples
- 17:11Challenges and potential false positives
- 18:37Results and impact on EasyList
- 20:25Acknowledging contributors and inspiration
- 21:06Manual review and site selection process
- 22:32Workflow and automation details
- 23:13Impact on filter list authors' workload
- 24:57Security concerns and recipe secrecy
- 26:40Feature requests and script capabilities
- 28:04Q&A session begins
- 28:27Denial modifier and EasyList rules
- 29:11Development time and challenges
- 30:36Holiday testing and results
- 31:05Ad scripts detecting automation
- 32:54Final thoughts and Q&A wrap-up
related talks
53mWait, I have to test the front end too?
51mAutomating to augment testing
18mAutomating All the Code & Testing Things with GitHub Actions
14mDebugging Secrets of a Lazy Developer
46mMagic of Automation and Everyday Chores
39mAutomating Your Security Checks: Stop Hiding Your Flaws Next to the Monsters...
01h 01mWebcrawler
40mHow To (Semi-)Automate JavaScript Refactoring