# vLLM allows Remote Code Execution by Pickle Deserialization via AsyncEngineRPCServer() RPC server entrypoints

Advisory record · `GHSA-cj47-qj6g-x7r4`

- **Identifiers:** `GHSA-cj47-qj6g-x7r4` · `CVE-2024-9053` · `PYSEC-2025-222`
- **State:** Confirmed
- **Severity:** CVSS 9.8 Critical
- **Critical:** Yes: CVSS 9+
- **Weakness (CWE):** [CWE-78](https://cwe.mitre.org/data/definitions/78.html) · [CWE-502](https://cwe.mitre.org/data/definitions/502.html)
- **Tags:** [Command injection](https://agentprime.io/advisories/tag/command-injection.md) · [Insecure deserialization](https://agentprime.io/advisories/tag/deserialization.md)
- **Published:** 20 Mar 2025
- **Updated:** 7 Aug 2026
- **Reviewed by a person:** No

Published by the GitHub Advisory Database at 20 Mar 2025, 12:32 UTC. agentprime has not analysed this yet, so whether it affects any app is not known.

## Critical

Critical, because:

- CVSS score 9.8: 9.0 or more.

How we decide what’s critical: <https://agentprime.io/advisories/critical.md>

## Affected packages

- `vllm` (PyPI): affected `<=0.6.0`; no fixed version yet

## Machine-written summary

Summary written by Claude Haiku from the source advisory. Not reviewed by a person.

> vLLM versions up to and including 0.6.0 are vulnerable to remote code execution through pickle deserialization in the AsyncEngineRPCServer RPC server entrypoints, where received messages are passed to cloudpickle.loads() without sanitization. An attacker who can send messages to the RPC server can supply malicious pickle data and run arbitrary code on the host. No fixed version is listed.

Written 5 Nov 2026.

## Not analysed yet

We haven’t published our analysis of this Advisory yet. Once a person has reviewed it, this page will say what’s vulnerable, what has to be true for your app to be affected, and how to check.

## Severity and exploitation

- 9.8 Critical, CVSS 3.0: `CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H`, from CNA (@huntr_ai)

- **CISA KEV:** Not listed.
- **EPSS:** 0.014, the 72nd percentile, on 17 Nov 2026: FIRST’s estimate of the chance of exploitation in the next 30 days.

## Timeline

- 31 Oct 2026, 14:40 UTC: First seen by agentprime
- 31 Oct 2026, 14:40 UTC: [Confirmed by OSV](https://osv.dev/vulnerability/GHSA-cj47-qj6g-x7r4)

## Sources

- [GitHub Advisory Database: GHSA-cj47-qj6g-x7r4](https://github.com/advisories/GHSA-cj47-qj6g-x7r4)
- [OSV: GHSA-cj47-qj6g-x7r4](https://osv.dev/vulnerability/GHSA-cj47-qj6g-x7r4)
- <https://nvd.nist.gov/vuln/detail/CVE-2024-9053>
- <https://github.com/pypa/advisory-database/tree/main/vulns/vllm/PYSEC-2025-222.yaml>
- <https://github.com/vllm-project/vllm>
- <https://huntr.com/bounties/75a544f3-34a3-4da0-b5a3-1495cb031e09>

Facts on this page come from the GitHub Advisory Database (CC-BY 4.0), OSV and the other sources listed, as published. We don’t reproduce full advisory descriptions, which can contain exploit code.

## Check whether this affects you

A Match means an installed version is in the affected range above. Most don’t affect the app; agentprime hasn’t analysed this one yet.

- With the agentprime MCP server (<https://api.agentprime.io/mcp>), ask your agent to check the project: `check_lockfile` sends a lockfile, `check_packages` a list of installed packages, and both return the Matches and the upgrades that clear them. `get_advisory` returns this record. Setup: <https://agentprime.io/mcp.md>
- In a browser: <https://agentprime.io/check/> reads your lockfile in the browser, never uploading it.
- Whether it touches your code, with file and line: the Claude Code plugin, <https://agentprime.io/plugin.md>
- As JSON: <https://api.agentprime.io/v1/advisories/GHSA-cj47-qj6g-x7r4>, or `POST https://api.agentprime.io/v1/match` with your packages (<https://agentprime.io/api.md>)

---

The web page: <https://agentprime.io/advisories/GHSA-cj47-qj6g-x7r4/>. Everything for agents: <https://agentprime.io/llms.txt>.
