Auto Browser Sniffer (AB Sniffer) Challenge
Overview
The AB Sniffer challenge tests miners' ability to detect and correctly identify browser automation frameworks. It evaluates page-local automation signals, framework-specific leaks, headless execution, and false positives in normal human-driven browsers.
Miners must maintain reliable classification across headed and headless execution modes while ensuring that human sessions do not trigger any automation or headless detector.
General Technical Requirements
- Development Language: Node.js SDK development
- Operating System: Ubuntu 24.04
- Environment: Docker container environment
- Architecture: amd64 (ARM64 at your own risk)
General Guidelines
- Detection Method: Analyze automation behavior, unique signatures, or behavioral patterns.
- Execution Modes: Detectors are tested in both headed and headless modes.
- Human Safety: A normal headed browser must not trigger a framework or headless detector.
- Prohibited Method: Browser fingerprinting is not allowed.
Plagiarism Check
We maintain strict originality standards:
- All submissions are compared against other miners' SDKs.
- 100% similarity = zero score.
- Similarity above 60% will result in rejection of the submission.
Submission Path
Dedicated Path: examples/miner_commit/src/commit/
Place your detection module files in this directory before building your commit:
nodriver.jsseleniumbase.jsseleniumdriverless.jspatchright.jspuppeteerextra.jszendriver.jsbotasaurus.jspydoll.jsheadless.js
Challenge Versions
Current:
- v6 - Framework detection, independent headless classification, and human verification
Deprecated:
Resources & Guides
- Building a Submission Commit - General submission instructions
- AB Sniffer v6 Testing Manual - Local validation and environment guidance
- Challenge Repository
- Miner Repository
📑 References
- Docker - https://docs.docker.com