Unverified50% confidenceReleaseExact time
HuggingFace has launched the text generation model GLM-5.3-CyberSecurity-FP8, built specifically for offensive security with refusal mechanisms removed, and officially labeled as suitable for use in authorized testing environments.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
UnverifiedHugging Face已采取SafeTensors格式替代不安全的Pickle序列化，并部署了恶意文件扫描机制76% similarVerifiedWhen Hugging Face used U.S. commercial model APIs for forensic analysis, the models' safety guardrails rejected requests involving exploit code and attack artifacts.73% similarUnverifiedHuggingFace has faced security challenges related to malicious models, supply chain security, and model persistence on multiple occasions.70% similarVerifiedSafeTensors是Hugging Face推出的安全序列化格式，支持零拷贝内存映射加载，并消除反序列化时的任意代码执行安全风险66% similarUnverifiedBetween 2023 and 2024, security researchers discovered malicious models on the Hugging Face platform multiple times, with attackers executing attacks by embedding malicious code in model serialization files.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/893012API
curl https://kongchang.com/api/v1/knowledge/claims/893012MCP
get_claim(id=893012)