Security Digest · Vulnerability record

CVE-2025-47277: vLLM Allows Remote Code Execution via PyNcclPipe Communication Service

vLLM, an inference and serving engine for large language models (LLMs), has an issue in versions 0.6.5 through 0.8.4 that ONLY impacts environments using the PyNcclPipe KV cache transfer integration with the V0 engine.

CVSS 9.8 Critical EPSS 1.0% · p61

Should you act?

Critical and remotely reachable

No known exploitation yet, but it is network-reachable without credentials. This signal is computed from CISA KEV, EPSS and CVSS data; it is not an editorial verdict.

  • No confirmed exploitation. It is not in CISA’s KEV catalog and CISA’s SSVC assessment records no known exploitation. That can change quickly after a public write-up; re-check if you are affected.
  • EPSS 1.0%, percentile 61: FIRST’s model estimates a 1.0% chance of exploitation activity in the next 30 days, higher than 61% of all scored CVEs (moderate). EPSS is a probability, not a verdict.
  • Remotely reachable with no login and no user interaction. If the vulnerable service is exposed to the internet, prioritize it.
  • Only applies if your dependency tree includes vllm (PyPI) >= 0.6.5, < 0.8.5, fixed in 0.8.5. Check your lockfile, including transitive dependencies.

What it is

vLLM, an inference and serving engine for large language models (LLMs), has an issue in versions 0.6.5 through 0.8.4 that ONLY impacts environments using the PyNcclPipe KV cache transfer integration with the V0 engine. No other configurations are affected. vLLM supports the use of the PyNcclPipe class to establish a peer-to-peer communication domain for data transmission between distributed nodes. The GPU-side KV-Cache transmission is implemented through the PyNcclCommunicator class, while CPU-side control message passing is handled via the sendobj and recvobj methods on the CPU side.​ The intention was that this interface should only be exposed to a private network using the IP address specified by the --kv-ip CLI parameter. The vLLM documentation covers how this must be limited to a secured network. The default and intentional behavior from PyTorch is that the TCPStore interface listens on ALL interfaces, regardless of what IP address is provided. The IP address given was only used as a client-side address to use. vLLM was fixed to use a workaround to force the TCPStore instance to bind its socket to a specified private interface. As of version 0.8.5, vLLM limits the TCPStore socket to the private interface as configured.

From the GitHub advisory

Impacted Environments

This issue ONLY impacts environments using the PyNcclPipe KV cache transfer integration with the V0 engine. No other configurations are affected.

Summary vLLM supports the use of the PyNcclPipe class to establish a peer-to-peer communication domain for data transmission between distributed nodes. The GPU-side KV-Cache transmission is implemented through the PyNcclCommunicator class, while CPU-side control message passing is handled via the sendobj and recvobj methods on the CPU side.​

A remote code execution vulnerability exists in the PyNcclPipe service. Attackers can exploit this by sending malicious serialized data to gain server control privileges.

The intention was that this interface should only be exposed to a private network using the IP address specified by the --kv-ip CLI parameter. The vLLM documentation covers how this must be limited to a secured network: https://docs.vllm.ai/en/latest/deployment/security.html

Weakness type: CWE-502.

Product: vllm (PyPI) · Assigned by: GitHub_M · Published: 2025-05-20

Affected and fixed versions

EcosystemPackageVulnerableFirst fixed
PyPIvllm>= 0.6.5, < 0.8.50.8.5
VendorProductAffected versions (per CVE record)
vllm-projectvllm>= 0.6.5, < 0.8.5

Severity scores

Scored byCVSSScoreVector
GitHub_M3.19.8 CriticalCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
GitHub3.19.8 CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
  • Attack vector: network (remotely reachable)
  • Attack complexity: low
  • Privileges required: none (no login)
  • User interaction: none
  • Scope: unchanged

CVSS measures how bad a flaw is if exploited, not how likely exploitation is. Read it next to the exploitation data below.

Exploitation and urgency

  • CISA KEV: not listed (no confirmed in-the-wild exploitation known to CISA).
  • CISA SSVC assessment: exploitation none, automatable yes, technical impact total.
  • FIRST EPSS (model of 2026-10-21): 0.96% probability of exploitation activity in the next 30 days; percentile 61.

Timeline

  • — CVE record last updated
  • — CVE record published by GitHub_M
  • — GitHub-reviewed advisory GHSA-hjq4-87xh-g4fv published (critical)

Sources and references

Original sources first. Version and fix details should always be confirmed against the vendor or project advisory.

Vulnerability records

  1. GitHub Advisory GHSA-hjq4-87xh-g4fv github.com
  2. CVE.org record CVE-2025-47277 cve.org
  3. NVD entry CVE-2025-47277 nvd.nist.gov

Other references from the CVE record and advisory

  1. github.com github.com
  2. github.com github.com
  3. github.com github.com
  4. docs.vllm.ai docs.vllm.ai
  5. github.com github.com
  6. github.com github.com
  7. pypi.org pypi.org

Critical PyPI advisories · Today’s Security Digest