Your RSA-2048 keys break in 2030. Find every one of them before attackers do.
🐍 PyPI

CVE-2026-54236

MEDIUM

vLLM: incomplete CVE-2026-22778 fix leaks PIL repr addresses via Anthropic router

Also known asGHSA-hgg8-fqqc-vfmwPYSEC-2026-3408
Published
Jun 22, 2026
Updated
Jul 15, 2026
Affected
1 pkg
Patched
1 / 1
Exploits
None indexed

Blast Radius

1 pkg affected
🐍vllm

Real-time download stats are indexed for npm and PyPI packages. This vulnerability affects PyPI packages — download data is not available via public APIs for these ecosystems.

Description

vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.23.1rc0, the fix for CVE-2026-22778, which introduced a sanitize_message helper that strips object-repr memory addresses from error messages before they reach the client, is incomplete: several response paths echo str(exc) directly to clients without calling sanitize_message. The unsanitized sites include the Anthropic API router in vllm/entrypoints/anthropic/api_router.py (the POST /v1/messages and POST /v1/messages/count_tokens handlers), the Server-Sent Events streaming converter in vllm/entrypoints/anthropic/serving.py, and the realtime speech-to-text WebSocket in vllm/entrypoints/speech_to_text/realtime/connection.py. These paths catch the exception inside the route coroutine and construct the JSONResponse themselves, bypassing the sanitizing global FastAPI exception handler, and WebSocket frames do not traverse that handler chain at all. Using the same primitive as the parent issue, an unauthenticated attacker can send malformed image bytes through the Anthropic Messages API image content parts so that PIL.Image.open raises an UnidentifiedImageError whose message contains the BytesIO object repr, leaking the heap memory address verbatim in the error.message field of the response body. This vulnerability is fixed in 0.23.1rc0.

Affected Packages

1 total 1 fixed
EcosystemPackageVulnerable rangeFix
🐍PyPIvllmall versions0.24.0

Detection & mitigation playbook

Open-source dependency
  1. Detect

    Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for vllm. O3's reachability analysis confirms whether the vulnerable code path is actually invoked in your application, so you act on real exposure instead of every transitive match.

  2. Fix

    Update vllm to 0.24.0 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms CVE-2026-54236 is resolved across your whole dependency graph.

  3. Workarounds

    If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.

  4. How O3 protects you

    O3 pinpoints whether CVE-2026-54236 is reachable in your code and exactly where to fix it, then blocks exploitation in production at runtime until the patched version is deployed.

Tailored to CVE-2026-54236. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.

Frequently Asked Questions

vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.23.1rc0, the fix for CVE-2026-22778, which introduced a sanitize_message helper that strips object-repr memory addresses from error messages before they reach the client, is incomplete: several response paths echo str(exc) directly to clients without calling sanitize_message. The unsanitized sites include the Anthropic API router in vllm/entrypoints/anthropic/api_router.py (the POST /v1/messages and POST /v1/messages/count_tokens handlers), the Server-Sent Events streaming converter in vllm/entrypoints/anthrop
O3 Security · Impact-Aware SCA

Is CVE-2026-54236 in your dependencies?

O3 detects CVE-2026-54236 across PyPI dependencies and uses function-level reachability to confirm whether the vulnerable code path is actually reachable — not just present. No false positives.