Out-of-bounds write in parsing IFD tag in libimagecodec.quram.so prior to SMR Dec-2025 Release 1 allows remote attackers to access out-of-bounds memory.
Out-of-bounds write in decoding metadata in fingerprint trustlet prior to SMR Dec-2025 Release 1 allows local privileged attackers to write out-of-bounds memory.
Improper export of android application components in Dynamic Lockscreen prior to SMR Dec-2025 Release 1 allows local attackers to access files with Dynamic Lockscreen's privilege.
vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.11.1, vllm has a critical remote code execution vector in a config class named Nemotron_Nano_VL_Config. When vllm loads a model config that contains an auto_map entry, the config class resolves that mapping with get_class_from_dynamic_module(...) and immediately instantiates the returned class. This fetches and executes Python from the remote repository referenced in the auto_map string. Crucially, this happens even when the caller explicitly sets trust_remote_code=False in vllm.transformers_utils.config.get_config. In practice, an attacker can publish a benign-looking frontend repo whose config.json points via auto_map to a separate malicious backend repo; loading the frontend will silently run the backend’s code on the victim host. This vulnerability is fixed in 0.11.1.
ChurchCRM is an open-source church management system. In ChurchCRM 6.2.0 and earlier, there is a time-based blind SQL injection in the handling of the 1FieldSec parameter. Injecting SLEEP() causes deterministic server-side delays, proving the value is incorporated into a SQL query without proper parameterization. The issue allows data exfiltration and modification via blind techniques.