Xinference
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 2
- Load/Mine Targets AI Users Through Fake Qwen2 Models
Pluto Security reported that the Load/Mine campaign distributed malicious repositories disguised as Qwen2 test models through Hugging Face, abusing the trust_remote_code=True option in Transformers. Researchers tracked 31 malicious repositories across seven Hugging Face accounts, 13 related models and 32,376 downloads; malicious instructions could run before model weights loaded without modifying the weights. After the first wave was removed, attackers added 19 repositories containing mining code, reverse shells and container-escape code, targeting self-hosted inference systems, GPU hosts, research notebooks and CI environments.
iThome 台湾 · 🔥 0 - Pluto Security details a trust_remote_code cryptomining campaign
Pluto Security’s analysis, reported by The Next Web, describes a cryptomining campaign on the Hugging Face Hub that executed Python code through the trust_remote_code=True model-loading path rather than an exploit. Malicious repositories presented themselves as small Qwen2 models and used imported Python modules to install persistence, open a reverse shell, download a disguised miner and report its status. Pluto Security found that repository scanners marked the plain Python payload as safe because it did not match a known malware signature or serialized-model format.
The Next Web · 🔥 0
Experience and discussion from the community
Share my Xinference experienceAsk about Xinference
Nobody has shared their experience with Xinference yet.