vLLM: Processing differential in multi-channel audio downmixing enables hidden-input/moderation bypass for audio models
PyPI vllm · Upgrade to 0.18.0 or later · >= 0.5.5, < 0.18.0 · CVE-2026-34760 · GHSA-6c4r-fmh3-7rh8 · amended
Published open-source security advisories, matched to the exact versions you run.
Run in a terminal
claude mcp add --scope user --transport http patchnext https://mcp.patchnext.com && claude mcp login patchnext
Your browser opens: sign in and choose what the agent may read.
Then ask: What do I need to patch on this machine?
26–26 of 26 matching · newest first
PyPI vllm · Upgrade to 0.18.0 or later · >= 0.5.5, < 0.18.0 · CVE-2026-34760 · GHSA-6c4r-fmh3-7rh8 · amended
A vendor fix does not resolve your finding until you verify it in your own systems.
Security reports that are not matched to your stack.
Listed as news: the source report lacks affected versions, a fix, or verified evidence.
Listed as news: the source report lacks affected versions, a fix, or verified evidence.
Listed as news: the source report lacks affected versions, a fix, or verified evidence.
Listed as news: the source report lacks affected versions, a fix, or verified evidence.
Listed as news: the source report lacks affected versions, a fix, or verified evidence.