Hacker Newsnew | past | comments | ask | show | jobs | submit | TimTom89's commentslogin

After losing €12k to invoice fraud in my German e-commerce business, I built a plugin that analyzes WooCommerce orders for risk patterns

The problem: In Germany, "pay by invoice" (Rechnung) is still huge - customers expect it. But combine that with packstation deliveries and fake addresses, you're bleeding money.

What it does: - Analyzes order patterns - Scores risk based on multiple factors - Optional API for enhanced detection (hosted in Germany for GDPR)

Just launched on WordPress.org. The basic version is free forever, planning to add premium features based on feedback.

Tech stack is simple: PHP/MySQL, optional REST API for enhanced checks.

Would love feedback from anyone dealing with e-commerce fraud, especially in markets where post-payment is common.


Open Scraping Challenge – Test our ML Bot Detection on Aegilock!

Hi all,

We openly invite anyone interested in scraping or security testing to try and scrape our production website, especially the login area:

https://www.aegilock.de/

We are running a 100% GDPR-compliant, fully ML-driven bot detection and logging engine. You are free to use your favorite scrapers, headless browsers, Selenium, Puppeteer, curl, Python scripts – anything goes! Feel free to target our login form, API endpoints, or other resources.

Why participate?

Help us benchmark and improve our ML-based detection

Get a free anonymized sample of the resulting log/events for your own research or ML training

What do we log?

No IPs or personal data – only anonymized behavioral & technical features, ML score, block status, and interaction vectors.

Interested? Just give it a try – and send us your feedback, attack logs, or findings!

Contact for cooperation, log samples, or partnership: kontakt@aegilock.de

Let’s make bot protection measurable and transparent together!

Best regards, Thomas Röhrig Founder, Aegilock


I just open-sourced a lightweight ML-based bot detection API. It uses simple features like user-agent entropy, path entropy, outdated browser patterns and sensitive URLs to detect automated requests.

Built with LightGBM and Flask No IP tracking, no cookies Ideal for edge use cases or form validation Easily extensible via feature hooks Model can be retrained from logs (retrain_model.py)

https://github.com/TimTom32hh/ml-bot-score-api

Would love feedback or contributors interested in privacy-first bot defense.


Hi HN,

I created *Aegilock*, an open-source, self-hosted bot protection solution designed as a GDPR-friendly alternative to traditional CAPTCHAs. It's invisible to users and doesn’t use cookies, US-based cloud services, or tracking.

It's based on a combination of Proof-of-Work challenges and server-side Machine Learning (ML) scoring, built with Node.js and Express.

I'd appreciate any feedback! You can check out the project here: https://github.com/DEIN_GITHUB_NAME/aegilock-onprem

Looking forward to your thoughts!


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: