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.
CVE-2026-54236
“—” beside the score: it is equal to the CVSS score beside it.
This medium-severity CVE scores 5.3 under a secondary CVSS source (NVD's own analysis pending). EPSS exploit probability: 0.9% (60th percentile of EPSS-scored CVEs). No GitHub Security Advisory CVSS is on record for this CVE.
Also known asEUVD-2026-38403 (ENISA EUVD) · GHSA-hgg8-fqqc-vfmw (GitHub Security Advisories)Source: ENISA EUVD
- CISA's Vulnrichment rates exploitation as PoC; EchelonGraph holds no exploit-code reference for it
A fix is available — apply it.
- CVSS v3
- 5.3
- EG Score
- 5.3MEDIUMmedium confidence
- EG Risk
- 55EG Risk 55/100
EG Risk is EchelonGraph's 0–100 priority score. It fuses impact severity (EG Severity, the CVSS base score, never the KEV-floored or EPSS-lifted EG Score) with real-world exploitation and automatability, under weights EchelonGraph set by hand, so you can rank equal-severity CVEs and fix the most dangerous first. Higher = act sooner.
How it’s computed (value × EchelonGraph’s weight)Severity53% × 45%Impact severity (EG Severity): CVSS 5.3 of 10. Never the KEV-floored or EPSS-lifted EG Score.Exploitation40% × 40%EchelonGraph's fixed weight for EG-KEV Tier 3 (Elevated): a constant, not a probability.Automatability100% × 15%CISA's SSVC Automatable value: yes (Vulnrichment).SSVC (CISA's decision table, applied by EchelonGraph): Track at low or medium mission impact; Attend at high (mission-essential systems).Exploitation public PoC (CISA Vulnrichment) · Automatable yes (CISA Vulnrichment) · Technical impact partial (CISA Vulnrichment). All three are CISA's SSVC values (Vulnrichment). EchelonGraph applied CISA's decision table to them; CISA publishes SSVC inputs and the table, not a decision for each CVE.Action: A fix is available. Apply it within your standard update timelines at low or medium mission impact and sooner than that at high. - EPSS PROB
- 0.9%
- EPSS %ILE
- 60th
- KEV
- Not listed
What the EG score is made of (scorer v4.0: the three questions the 0–10 score mixes, published apart · methodology)
The EG score (5.3): severity, escalated by confirmed exploitation. This CVE's 5.3 is its impact severity (EG Severity), not escalated.
- EG Severity
- 5.3 MEDIUMCVSS-B, impact only: no KEV listing, EPSS or EG-KEV tier raises it.a secondary CVSS source · CVSS v3.0 or v3.1
- EG Likelihood
- 0.9% within 30 daysSource: FIRST EPSS (model v2026.06.15, score date 2026-10-25). FIRST's exploitation probability, not an EchelonGraph model.
- Confirmed exploitation
- None on record (CISA KEV, EG-KEV Tier 1)Reported apart from the likelihood: an observed fact, not a probability.
- EG Priority
- SSVC Track at medium mission impactTrack at low or medium mission impact; Attend at high (mission-essential systems).CISA BOD 26-04: 14 days if the affected system is publicly exposed; 60 days if it is not.Each input and its source, and whom BOD 26-04 binds: the SSVC and BOD 26-04 blocks below.
SSVC (CISA's decision table, applied by EchelonGraph)Track at low or medium mission impact; Attend at high (mission-essential systems).
A fix is available. Apply it within your standard update timelines at low or medium mission impact and sooner than that at high.
Exploitation public PoC (CISA Vulnrichment) · Automatable yes (CISA Vulnrichment) · Technical impact partial (CISA Vulnrichment). All three are CISA's SSVC values (Vulnrichment). EchelonGraph applied CISA's decision table to them; CISA publishes SSVC inputs and the table, not a decision for each CVE. Mission impact is CISA's Mission & Well-being decision point, and only you can judge it. CISA's table makes it high when Mission Prevalence is Essential — the vulnerable component "directly provides capabilities that constitute at least one MEF for at least one entity" (MEF: mission essential function) — or when Public Well-Being Impact is Irreversible: "multiple fatalities are likely", the cyber-physical system "is likely lost or destroyed", "extreme or serious externalities" are imposed on other parties, or social systems such as elections or the financial grid "are destabilized and potentially collapse". CISA's decision table
BOD 26-04 (CISA's remediation timeline, Table 1)14 days if the affected system is publicly exposed; 60 days if it is not.
The score's inputs: In CISA KEV no (CISA's KEV catalog) · Automatable yes (CISA Vulnrichment) · Technical impact partial (CISA Vulnrichment). Whether your system is publicly exposed is yours to answer. BOD 26-04 binds US Federal Civilian Executive Branch agencies; for anyone else it is CISA's published timeline, a reference rather than an obligation. The directive · how EchelonGraph applies it
Published
June 17, 2026
Last Modified
October 25, 2026
Advisory Details (3)
Auto-updated Jun 23, 2026commit 949236297293 (vllm-project/vllm)
Fix landed in vllm-project/vllm commit 949236297293 — awaiting tagged release
https://github.com/vllm-project/vllm/commit/94923629729381d7f7c9efde72071a2441f7fd82[Security] Apply sanitize_message to Anthropic and STT error paths
Fix merged in vllm-project/vllm PR #45119 on 2026-06-11 — awaiting tagged release
https://github.com/vllm-project/vllm/pull/45119vLLM: incomplete CVE-2026-22778 fix leaks PIL repr addresses via Anthropic router (CWE-532) · Advisory · vllm-project/vllm · GitHub
https://github.com/vllm-project/vllm/security/advisories/GHSA-hgg8-fqqc-vfmwVendor Advisories for CVE-2026-54236(1)
These vendors published their own advisory mentioning this CVE — often with vendor-specific remediation steps + affected product lists not in NVD.
Patch Availability(1)
| Vendor / Ecosystem | Fix (by version range) / Patch | Released | Source |
|---|---|---|---|
| pip | vllm | — | ghsa |
Patches are aggregated from vendor advisories (Red Hat, Microsoft, Cisco, GitHub) and package ecosystems (OSV, GHSA). Multiple rows for the same upstream release have been deduplicated.
Affected Packages
(1 across 1 ecosystem)
PyPI(1)
| Package | Vulnerable range | Fix by version range | Dependents |
|---|---|---|---|
| vllm | 0.0.1 ... 0.9.2 (89 versions) |
| — |
Weakness Classification(1)
MITRE Common Weakness Enumeration — the root-cause categories this CVE belongs to.
Data Freshness Timeline
(refreshed 7× in last 7d / 38× in last 30d)
Each row is a time one of our pipelines fetched, updated or re-scored this CVE. For example, "NVD update" means NVD published or revised its analysis for this CVE; "MITRE cvelistV5" means we ingested or refreshed it from the CNA feed. An arrow marks a change the record states ("CVSS v3 → 9.1"); EPSS and GHSA rows show the values recorded at that time, not what changed (EPSS values are rounded; hover one for the exact figures). Most recent first.
Showing the most recent 100 of 198 refreshes on record for this CVE.
A repeat, a row identical to one the same pipeline wrote less than 10 seconds earlier, is not a separate refresh and is not counted. None are among the latest 100 rows. The 98 oldest rows on record are not shown and were not checked, so the counts can still include repeats.
- 2026-11-02 16:45 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-11-01 16:34 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-31 16:33 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-30 16:34 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-29 16:33 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-28 16:32 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-27 16:32 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-26 16:31 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-25 23:37 UTCEG score recompute
- 2026-10-25 16:33 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-25 14:19 UTCEG score recompute
- 2026-10-24 16:33 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-23 16:32 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-22 21:23 UTCEG score recompute
- 2026-10-22 16:34 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-21 16:33 UTCEPSS rescoreEPSS 0.9% · 60th percentile
- 2026-10-20 16:33 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-19 16:32 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-18 16:32 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-18 03:31 UTCEG score recompute
- 2026-10-17 16:34 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-16 16:42 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-15 16:32 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-14 16:33 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-13 16:34 UTCEPSS rescoreEPSS 0.9% · 59th percentile
Show 75 moreShow fewer
- 2026-10-12 16:34 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-12 15:56 UTCEG score recompute
- 2026-10-11 19:39 UTCEG score recompute
- 2026-10-11 16:55 UTCEG score recompute
- 2026-10-11 14:18 UTCEG score recompute
- 2026-10-11 13:49 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-10 20:00 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-09 20:07 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-08 17:12 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-07 23:48 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-06 15:18 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-05 17:43 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-04 23:22 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-03 14:26 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-03 06:09 UTCOSV refresh
- 2026-10-02 17:56 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-10-01 19:50 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-09-30 15:03 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-09-29 15:22 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-28 13:51 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-27 13:48 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-26 15:59 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-24 14:04 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-23 17:54 UTCEPSS rescoreEPSS 0.9% · 59th percentile
- 2026-09-20 20:16 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-18 19:28 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-17 19:31 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-16 14:08 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-16 05:15 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-14 03:08 UTCEG score recompute
- 2026-09-13 16:47 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-11 14:53 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-10 09:35 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-08 22:01 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-07 16:01 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-06 13:47 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-04 05:06 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-02 14:12 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-01 13:54 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-09-01 04:40 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-08-30 22:26 UTCEG score recompute
- 2026-08-30 19:17 UTCEPSS rescoreEPSS 0.9% · 58th percentile
- 2026-08-30 01:22 UTCEPSS rescoreEPSS 0.9% · 57th percentile
- 2026-08-28 21:41 UTCEPSS rescoreEPSS 0.9% · 57th percentile
- 2026-08-27 14:25 UTCEPSS rescoreEPSS 0.9% · 57th percentile
- 2026-08-26 14:46 UTCEPSS rescoreEPSS 0.9% · 57th percentile
- 2026-08-25 13:49 UTCEPSS rescoreEPSS 0.9% · 57th percentile
- 2026-08-25 06:16 UTCEG score recompute
- 2026-08-23 04:29 UTCEG score recompute
- 2026-08-23 00:19 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-21 23:49 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-21 07:24 UTCEG score recompute
- 2026-08-20 22:55 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-20 08:16 UTCEG score recompute
- 2026-08-19 17:04 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-19 08:12 UTCEG score recompute
- 2026-08-18 13:48 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-17 15:27 UTCEG score recompute
- 2026-08-17 13:47 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-16 16:20 UTCEG score recompute
- 2026-08-16 14:56 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-16 02:14 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-15 17:17 UTCEG score recompute
- 2026-08-15 01:30 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-14 18:10 UTCEG score recompute
- 2026-08-13 22:00 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-12 17:54 UTCEG score recompute
- 2026-08-12 13:51 UTCEPSS rescoreEPSS 0.8% · 54th percentile
- 2026-08-11 18:52 UTCEG score recompute
- 2026-08-11 00:00 UTCEPSS rescoreEPSS 0.8% · 53rd percentile
- 2026-08-09 20:49 UTCEG score recompute
- 2026-08-09 13:46 UTCEPSS rescoreEPSS 0.8% · 53rd percentile
- 2026-08-08 21:48 UTCEG score recompute
- 2026-08-08 16:37 UTCEPSS rescoreEPSS 0.8% · 53rd percentile
- 2026-08-07 22:46 UTCEG score recompute
Public detection templates
(1 reference)A public Nuclei detection template is referenced for this CVE. A detection template checks whether a system is affected; it is not necessarily exploit code.
- Nucleihttp/cves/2026/CVE-2026-54236.yamlPublished Jun 22, 2026
vLLM <= 0.23.0 - Anthropic Router Heap Address Information Leak
Open source ↗
Related CVEs(same product + same CWE)
Same product
10 shownPyPI:vllm
Frequently asked(5)
What is CVE-2026-54236?
When was the CVE record for CVE-2026-54236 published?
Is CVE-2026-54236 actively exploited?
What is the CVSS score of CVE-2026-54236?
How do I remediate CVE-2026-54236?
Dependency Blast Radius
See which npm, PyPI, Go, and Maven packages are affected by CVE-2026-54236
Is Your Infrastructure Affected by CVE-2026-54236?
EchelonGraph automatically scans your cloud infrastructure and maps CVE exposure using blast radius analysis.