Zero-Day & Vulnerability Intelligence Hub
Echtzeit-Tracking mit EPSS Exploit-Wahrscheinlichkeiten, Angriffsvektor-Decodern und KI-Patch-Anleitungen.
📊 Historien-Charts — Criticals-Trend · Vendors · EPSS-Verteilung
| Tier | CVEs | Anteil |
|---|---|---|
| ≥90 % | 750 | 0,2 % |
| ≥50 % | 3.272 | 0,8 % |
| ≥10 % | 22.091 | 5,7 % |
| <10 % | 358.955 | 93,2 % |
CVE-2026-115912 | vLLM is an inference and serving engine for large language models. Prior to 0.32.0, an authenticated caller can place forged shared-memory address and monotonic_id values in public multimodal features.kwargs_data. In scale-out deployments with shared-memory multimodal caching enabled, workers trust those values and pass attacker-selected bytes to pickle.loads(), allowing arbitrary Python execution even when insecure serialization is disabled. This issue is f
vLLM is an inference and serving engine for large language models. Prior to 0.32.0, an authenticated caller can place forged shared-memory address and monotonic_id values in public multimodal features.kwargs_data. In scale-out deployments w
CVE-2026-115190 | An insecure deserialization (PHP object injection) vulnerability exists in Mautic's handling of segment filters and custom field properties. Several code paths that read these stored values from the database deserialize them without restricting which PHP classes may be instantiated, instead of using Mautic's safe deserialization helper. An attacker who is able to store a crafted serialized value in these segment filter or custom field properties can cause ar
An insecure deserialization (PHP object injection) vulnerability exists in Mautic's handling of segment filters and custom field properties. Several code paths that read these stored values from the database deserialize them without restric
CVE-2026-57845 | Hugging Face Datasets through 5.1.0 contains an insecure deserialization vulnerability that allows attackers to execute arbitrary Python code by crafting a malicious pickle file loaded through the deprecated pandas packaged builder. Attackers can supply a pickle-backed dataset that is processed by Pandas._generate_tables via pandas.read_pickle() without validation, so when a victim calls datasets.load_dataset("pandas", data_files=...), the embedded __reduce__
Hugging Face Datasets through 5.1.0 contains an insecure deserialization vulnerability that allows attackers to execute arbitrary Python code by crafting a malicious pickle file loaded through the deprecated pandas packaged builder. Attacke