dChan
Anonymous ID: a4ebc9 Aug. 5, 2026, 3:51 p.m. No.24906465   🗄️.is 🔗kun

Palo Alto Unit 42 (Their research team) just published that a Chinese attacker built himself an autonomous attack machine out of DeepSeek and Hermes agent - (OSS agent that creates skills based on successful tasks and becomes smarter by the minute).

He ran it in YOLO mode where the agent executes every command, even a dangerous one, without asking anyone for approval.

He wired it to a Telegram channel to receive instructions, gave it offensive skills, and connected it to the search engine.

The agent scanned GitHub for exploits by star count, and once it picked one, it wrote: "This one's for n8n, 258 stars and CVSS 10.0 — looks insanely promising."

It found 647,017 exposed n8n machines around the world, 25,209 of them in China.

It located three machines running a vulnerable version, and then wrote to itself: "Hold on I re-read the exploit code, it needs a form without authentication. These forms have authentication. Moving on to other targets."

And how did the folks at Unit 42 catch it?

The agent got a command over Telegram to spin up a file server and spun it up from the home directory instead of an isolated one, exposing the attacker's entire working environment.

 

All the reconnaissance, target selection, vulnerability research and scope narrowing. Dozens of hours of manual attacker work happened in minutes.

 

https://www.linkedin.com/posts/omer-talmy-ai101_palo-alto-unit-42-their-research-team-just-share-7490313322823618560-5gwq/

Anonymous ID: a4ebc9 Aug. 5, 2026, 3:55 p.m. No.24906470   🗄️.is 🔗kun

London cops handed victim's new address and number to her stalker, watchdog says

 

Met ordered to improve safeguards after two preventable data breaches

 

https://www.theregister.com/security/2026/08/05/london-cops-handed-victims-new-address-and-number-to-her-stalker-watchdog-says/5283382

Anonymous ID: a4ebc9 Aug. 5, 2026, 3:57 p.m. No.24906472   🗄️.is 🔗kun

Time Magazine has a separate version of its website with ads only AI can see

 

Brands are already paying to influence what chatbots say about them

 

AI webpage crawlers are now being served their very own ads that ordinary visitors never see, with one firm's boss openly describing a strategy to influence what chatbots say about brands. The next time you ask Claude about where to bank, its answer may have been influenced by this bot-targeted content.

 

https://www.theregister.com/ai-and-ml/2026/08/05/time-magazine-has-a-separate-version-of-its-website-with-ads-only-ai-can-see/5283640

 

seems appropriate (vidrel)