6 HIGH · 1 INFO · THREAT ORANGE · 7 items · Generated in 415s
A vulnerability in langchain-ai langgraph up to 1.2.4 allows remote attacks with high complexity and difficulty.
vLLM allocates memory proportional to uploaded file size, causing memory pressure or process termination.
A remote client can crash the shared engine worker, causing a service-wide denial of service for other clients.
A vulnerability in the vLLM library allows remote users to induce a server application crash.
A high-throughput and memory-efficient inference engine for LLMs has a denial of service vulnerability due to unvalidated regex patterns.
Unauthenticated command execution vulnerability in flyto-core allows attackers to execute arbitrary OS commands as the server process, posing a high-severity lo
A challenge to hack an AI assistant was run, with 6,000 attempts failing to leak secrets.