101010.pl is one of the many independent Mastodon servers you can use to participate in the fediverse.
101010.pl czyli najstarszy polski serwer Mastodon. Posiadamy wpisy do 2048 znaków.

Server stats:

504
active users

#darkweb

0 posts0 participants0 posts today

This week's Top 5:

➡️ Dark-web dealing
➡️ A cruel tradition
➡️ Revisiting Twain
➡️ Flushing in flux
➡️ An unlikely reunion

Our editors recommend excellent #longreads this week by Andy Greenberg, Michelle Orange, John Jeremiah Sullivan, Jefferson Mao, and Will Steinfeld.

longreads.com/2025/05/30/the-t

Longreads · The Top 5 Longreads of the WeekBy Longreads

AIL 6.2 released - Smarter Analysis, Search and Enhanced User Experience

We’re excited to release AIL Framework v6.2, a major update with new features and improved performance. This version makes analysis easier and the overall experience faster and more user-friendly.

Among the highlights are a fully revamped search engine powered by MeiliSearch, improved language detection for short text, local AI-driven image descriptions, and a yara-hunting editor tool.

🔗 ail-project.org/blog/2025/05/2

We implemented a major new feature in the AIL Project that addresses a long-standing issue related to the collection of images or screenshots that may be harmful to analysts (e.g., violent content, CSAM, etc.). The feature allows users to trigger the description of an image before actually viewing it.

The feature will be included in the upcoming release of AIL (version 6.2).

This work is co-funded in the AIPITCH project. We would like to thank Qwen for the open source Qwen2-VL vision-language models which provide an excellent basis for image detection and description while allowing local inferences.

@aipitch @circl @ail_project

I had the pleasure of presenting at #FIRSTCTI25 in Berlin:

"The Art of Pivoting – How You Can Discover More from Adversaries with Existing Information."

The talk explored how unconventional indicators, like cookie names, QR codes, HTTP headers (HHHash), DOM structures, and reused Google Analytics IDs, can reveal surprising links across threat actor infrastructure and behavior.

We also shared real-world insights from our crawling and analysis with AIL, including:

  • How “weak” indicators can gain strength through composite correlation
  • Unexpected metadata reuse across Tor services and social platforms
  • How AIL enables more creative and effective pivoting workflows

🔗 Slides ail-project.org/assets/img/sli

#threatintel #threatintelligence #cti #opensource #cybersecurity #darkweb

@misp @ail_project @circl

Thanks to @terrtia for the crazy discussions around correlations!

⚠️ Major AI data breach: DeepSeek leaks 1M+ records to the Dark Web 🧠💥

China-based AI startup DeepSeek just exposed a staggering volume of sensitive data:
💬 1M+ chat logs
🔐 API keys, backend metadata
📡 Unencrypted traffic via iOS app (ATS disabled)
🔓 Open ClickHouse database with full control

This wasn’t just a lapse — it was a floodgate.

🛡️ AI companies (and anyone integrating LLMs) must:
🔍 Audit storage configs
🔐 Enforce secure transport policies
📊 Monitor for unintentional data exposure

The future of AI relies on building trust. That starts with securing it.

#AI #CyberSecurity #DataBreach #DarkWeb #InfoSec
darkreading.com/cyberattacks-d

🕵️ Prodaft is taking an aggressive new approach to threat intelligence:
They’re buying access to major dark web forums — including admin and moderator accounts — in exchange for cryptocurrency.

Here’s what they’re offering:
💸 Crypto payments, no questions asked
🔐 Full anonymity for sellers
📊 Access to five top-tier forums
🎯 Prioritized payouts for higher access levels

The move aims to give security teams unprecedented visibility into cybercriminal operations — but it also pushes ethical and operational boundaries.

Would your team use intel gathered this way?

#Cybersecurity #ThreatIntel #DarkWeb #InfoSec #Ethics

darkreading.com/threat-intelli