vllm — Vulnerabilities & Security Advisories 98

This page aggregates known security vulnerabilities affecting the vLLM software product, a high-throughput inference engine for large language models. It collects advisories related to specific weakness types, covering the period from the project's initial release through the most recent updates. Readers can use this resource to track the vendor's published advisories, understand common vulnerability classes within the codebase, and review the product's historical security record without hunting through individual commit messages or release notes. The collection focuses on flaws in input validation, resource management, and deserialization, reflecting the primary attack surfaces identified by the community. This summary provides a consolidated view of past issues to support risk assessment and patching strategies.