AI-Powered ‘OpenClaw Trap’ Lures Developers With Trojanized GitHub Repos

Netskope Threat Labs has uncovered a sprawling malware campaign dubbed “TroyDen’s Lure Factory,” which uses over 300 trojanized GitHub repositories to distribute a custom LuaJIT infostealer.

Threat actors lure victims by impersonating highly sought-after tools, most notably an OpenClaw AI deployer, as well as gaming cheats, Roblox scripts, and phone-tracking utilities.

By combining polished fake repositories with artificial social proof, the campaign successfully compromises a wide range of targets across different communities.

The attack highlights a growing trend of sophisticated malware infection vectors hosted directly on developer platforms.

Evasion Tactics and Two-Part Payload

The malware’s core strength lies in its sophisticated two-component design, specifically engineered to bypass automated security sandboxes. .

Because automated security scanners typically analyze files individually, both the runtime and the encrypted text file appear harmless on their own.

The malicious behavior only emerges when a victim runs the batch file, which commands the interpreter to execute the hidden payload.

Once launched, the malware immediately runs five anti-analysis checks. It scans for debuggers, low-memory environments, and brief system uptimes to determine if it is trapped inside a security sandbox.

If it suspects an analysis environment, the malware executes a sleep command programmed to last roughly 29,000 years, completely outlasting any security timing window.

Scalable Infrastructure and AI Generation

The massive scale of this malicious operation strongly suggests the use of automated, AI-assisted malware generation.

Instead of relying on human-curated names, the campaign’s subdirectories use highly obscure terms drawn from archaic Latin, extinct biological taxonomy, and rare medical conditions.

For example, the OpenClaw repository hides its payload inside a folder named after an extinct order of birds.

At the same time, other packages use complex medical terminology. This programmatic naming allows the attackers to maintain a massive volume of simultaneous campaigns without manual effort.

According to Netskope research, this identical malware toolchain has been confirmed across multiple distinct lures, including a phone number tracker promoted via a popular Telegram channel and a prominent fishing game cheat.

All of these variants communicate with the same highly scalable infrastructure. Security analysts discovered that the command server utilizes a single management panel hidden behind eight load-balanced IP addresses.

The server endpoints even display signs of AI-assisted coding, featuring simple API routes for receiving screenshots and distributing new malicious tasks.

With over 300 confirmed delivery packages targeting diverse audiences, this operation highlights a dangerous evolution in cybercrime.

Indicators Of Compromise

Follow us on Google News , LinkedIn and X to Get More Instant Updates. Set Cyberpress as a Preferred Source in Google.