CVE-2026-93838

MEDIUMPre-NVD 5.95.9
EchelonGraph scoreMEDIUM confidence

“—” beside the score: it is equal to the CVSS score beside it.

Score 5.9 from GitHub Security Advisory (severity: HIGH) published 2026-09-18. The CNA's CVSS baseline 5.9; sources differ by 0.0.

Triggered by: GitHub Security Advisory CVSS
Sources: the CNA's CVSS, EPSS, GitHub Security Advisory
Elevated

Also known asEUVD-2026-83282 (ENISA EUVD) · GHSA-2wcg-j54r-m6wm (GitHub Security Advisories)Source: ENISA EUVD

5.9EG
EchelonGraph verdictPlan mitigationNo fix is confirmed yet: check the vendor's advisory for a fixed release; until one is applied, restrict network exposure of the affected system or apply the vendor's mitigation on your normal schedule (the SSVC decision on this page).
  • CISA's Vulnrichment rates exploitation as PoC; EchelonGraph holds no exploit-code reference for it
CISA-KEV: Not listedEPSS PROB: 0.7%CVSS: 5.9Exploit: No code on recordExposed services: Not assessed

No fix is confirmed yet — apply a workaround or compensating control (WAF / firewall / segmentation) and watch for the fix.

The CVE record's CNA (VulnCheck) lists sgl-project sglang versions up to and including 0.5.20 as affected. No release containing a fix is confirmed on record.

SGLang versions through 0.5.20 contain an unbounded memory allocation vulnerability in handle_staging_req() that fails to validate chunk_idx from ZMQ STAGING_REQ frames in prefill/decode disaggregation deployments. Attackers with access to the decode engine's internal ZMQ rank port can send a frame with an extremely large chunk_idx value, causing the scheduler to allocate memory until the system runs out and terminates the process.

CVSS v3
5.9
EG Score
5.9MEDIUMmedium confidence
EG Risk
43
EG Risk 43/100

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)
Severity59% × 45%
Impact severity (EG Severity): CVSS 5.9 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.
Automatability0% × 15%
CISA's SSVC Automatable value: no (Vulnrichment).
SSVC (CISA's decision table, applied by EchelonGraph): Track at low or medium mission impact; Track* at high (mission-essential systems).
Exploitation public PoC (CISA Vulnrichment) · Automatable no (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: No fix is confirmed yet. Check the vendor's advisory for a fixed release; until one is applied, restrict network exposure of the affected system or apply the vendor's mitigation on your normal schedule.
EPSS PROB
0.7%
EPSS %ILE
51st
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 (5.9): severity, escalated by confirmed exploitation. This CVE's 5.9 is its impact severity (EG Severity), not escalated.

EG Severity
5.9 MEDIUMCVSS-B, impact only: no KEV listing, EPSS or EG-KEV tier raises it.GitHub Security Advisory CVSS
EG Likelihood
0.7% within 30 daysSource: FIRST EPSS (model v2026.06.15, score date 2026-10-24). 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; Track* at high (mission-essential systems).CISA BOD 26-04: 60 days if the affected system is publicly exposed; Fix on system upgrade if it is not.No released fix is confirmed on this page, but CISA's timeline does not wait for a fix's release: it starts when an agency identifies the flaw on an asset and updates the CDM Agency and Federal Dashboard, and the directive counts patching, decommissioning the system or another action as remediation. Until a fixed release is applied, mitigate as this page's verdict says. In the directive's words, “Fix on system upgrade” means that, “unless conditions change”, the vulnerability “should be remediated the next time the vulnerable asset receives a scheduled major upgrade or rebuild”: its timelines are “dynamic as facts change”, and a KEV listing shortens them.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; Track* at high (mission-essential systems).

No fix is confirmed yet. Check the vendor's advisory for a fixed release; until one is applied, restrict network exposure of the affected system or apply the vendor's mitigation on your normal schedule.

Exploitation public PoC (CISA Vulnrichment) · Automatable no (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)60 days if the affected system is publicly exposed; Fix on system upgrade if it is not.

No released fix is confirmed on this page, but CISA's timeline does not wait for a fix's release: it starts when an agency identifies the flaw on an asset and updates the CDM Agency and Federal Dashboard, and the directive counts patching, decommissioning the system or another action as remediation. Until a fixed release is applied, mitigate as this page's verdict says. In the directive's words, “Fix on system upgrade” means that, “unless conditions change”, the vulnerability “should be remediated the next time the vulnerable asset receives a scheduled major upgrade or rebuild”: its timelines are “dynamic as facts change”, and a KEV listing shortens them.

The score's inputs: In CISA KEV no (CISA's KEV catalog) · Automatable no (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

September 18, 2026

Last Modified

September 24, 2026

Advisory Details (6)

Auto-updated Sep 18, 2026
No fix on record yet.
Web page

SGLang through 0.5.20 Unbounded Memory Allocation via STAGING_REQ chunk_idx | Advisories | VulnCheck

https://www.vulncheck.com/advisories/sglang-through-0.5.20-unbounded-memory-allocation-via-staging-req-chunk-idx
Web page

[Bug] Unbounded chunk_idx in PD disaggregation staging path causes Decode scheduler OOM · Issue #39764 · sgl-project/sglang · GitHub

https://github.com/sgl-project/sglang/issues/39764
Web page

sglang/python/sglang/srt/disaggregation/common/staging_handler.py at v0.5.20 · sgl-project/sglang · GitHub

https://github.com/sgl-project/sglang/blob/v0.5.20/python/sglang/srt/disaggregation/common/staging_handler.py#L848-L860
Web page

sglang/python/sglang/srt/disaggregation/common/staging_handler.py at v0.5.20 · sgl-project/sglang · GitHub

https://github.com/sgl-project/sglang/blob/v0.5.20/python/sglang/srt/disaggregation/common/staging_handler.py#L781
Web page

sglang/python/sglang/srt/disaggregation/common/conn.py at v0.5.20 · sgl-project/sglang · GitHub

https://github.com/sgl-project/sglang/blob/v0.5.20/python/sglang/srt/disaggregation/common/conn.py#L1636-L1640
Web page

GitHub - sgl-project/sglang: SGLang is a high-performance serving framework for large language models and multimodal models. · GitHub

https://github.com/sgl-project/sglang

Vendor Advisories for CVE-2026-93838(1)

These vendors published their own advisory mentioning this CVE — often with vendor-specific remediation steps + affected product lists not in NVD.

Weakness Classification(1)

MITRE Common Weakness Enumeration — the root-cause categories this CVE belongs to.

Data Freshness Timeline

(refreshed 8× in last 7d / 55× 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.

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. 1 repeat row among the latest 95 is merged into the row it repeats.

  1. 2026-11-01 16:36 UTCEPSS rescoreEPSS 0.7% · 51st percentile
  2. 2026-10-31 16:33 UTCEPSS rescoreEPSS 0.7% · 51st percentile
  3. 2026-10-30 16:35 UTCEPSS rescoreEPSS 0.7% · 51st percentile
  4. 2026-10-29 16:34 UTCEPSS rescoreEPSS 0.7% · 51st percentile
  5. 2026-10-28 16:33 UTCEPSS rescoreEPSS 0.7% · 51st percentile
  6. 2026-10-28 07:25 UTCOSV refresh
  7. 2026-10-27 16:33 UTCEPSS rescoreEPSS 0.7% · 51st percentile
  8. 2026-10-26 16:33 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  9. 2026-10-25 16:34 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  10. 2026-10-25 13:49 UTCEG score recompute
  11. 2026-10-24 16:34 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  12. 2026-10-23 16:33 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  13. 2026-10-22 16:35 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  14. 2026-10-21 16:34 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  15. 2026-10-20 16:34 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  16. 2026-10-19 16:33 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  17. 2026-10-18 22:01 UTCEG score recompute
  18. 2026-10-18 22:00 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  19. 2026-10-18 16:33 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  20. 2026-10-18 02:58 UTCEG score recompute
  21. 2026-10-17 16:35 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  22. 2026-10-16 16:45 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  23. 2026-10-15 16:33 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  24. 2026-10-14 16:34 UTCEPSS rescoreEPSS 0.7% · 50th percentile
  25. 2026-10-13 16:35 UTCEPSS rescoreEPSS 0.7% · 49th percentile
Show 69 more
  1. 2026-10-12 16:35 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  2. 2026-10-12 14:42 UTCEG score recompute
  3. 2026-10-11 19:12 UTCEG score recompute
  4. 2026-10-11 16:28 UTCEG score recompute
  5. 2026-10-11 13:50 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  6. 2026-10-11 13:41 UTCEG score recompute
  7. 2026-10-11 08:00 UTCEG score recompute
  8. 2026-10-11 07:59 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  9. 2026-10-10 20:00 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  10. 2026-10-10 06:48 UTCEG score recompute
  11. 2026-10-10 06:48 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  12. 2026-10-09 20:08 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  13. 2026-10-09 05:45 UTCEG score recompute
  14. 2026-10-09 05:45 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  15. 2026-10-08 17:14 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  16. 2026-10-08 04:43 UTCEG score recompute
  17. 2026-10-08 04:43 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  18. 2026-10-07 23:48 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  19. 2026-10-07 03:40 UTCEG score recompute
  20. 2026-10-07 03:40 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  21. 2026-10-06 15:19 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  22. 2026-10-06 02:29 UTCEG score recompute
  23. 2026-10-06 02:29 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  24. 2026-10-05 17:44 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  25. 2026-10-05 01:26 UTCEG score recompute
  26. 2026-10-05 01:26 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  27. 2026-10-04 23:23 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  28. 2026-10-04 00:23 UTCEG score recompute
  29. 2026-10-04 00:23 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  30. 2026-10-03 14:26 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  31. 2026-10-02 23:20 UTCEG score recompute
  32. 2026-10-02 23:20 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  33. 2026-10-01 22:13 UTCEG score recompute
  34. 2026-10-01 22:13 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  35. 2026-10-01 19:51 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  36. 2026-09-30 21:10 UTCEG score recompute
  37. 2026-09-30 21:10 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  38. 2026-09-30 15:04 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  39. 2026-09-29 20:08 UTCEG score recompute
  40. 2026-09-29 20:07 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  41. 2026-09-28 19:05 UTCEG score recompute
  42. 2026-09-28 19:05 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  43. 2026-09-28 13:52 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  44. 2026-09-27 18:02 UTCEG score recompute
  45. 2026-09-27 18:02 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  46. 2026-09-27 13:49 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  47. 2026-09-26 16:59 UTCEG score recompute
  48. 2026-09-26 16:59 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  49. 2026-09-26 15:59 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  50. 2026-09-25 15:53 UTCEG score recompute
  51. 2026-09-25 15:53 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  52. 2026-09-24 14:51 UTCEG score recompute
  53. 2026-09-24 14:51 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  54. 2026-09-24 14:04 UTCEPSS rescoreEPSS 0.7% · 49th percentile
  55. 2026-09-23 21:55 UTCEG score recompute
  56. 2026-09-23 21:55 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  57. 2026-09-23 17:54 UTCEPSS rescoreEPSS 0.5% · 38th percentile
  58. 2026-09-22 20:52 UTCEG score recompute
  59. 2026-09-22 20:52 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  60. 2026-09-22 16:01 UTCEPSS rescoreEPSS 0.5% · 38th percentile
  61. 2026-09-22 00:36 UTCEG score recompute
  62. 2026-09-22 00:35 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  63. 2026-09-21 21:09 UTCEPSS rescoreEPSS 0.5% · 38th percentile
  64. 2026-09-20 23:33 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  65. 2026-09-19 22:30 UTCEG score recompute
  66. 2026-09-19 22:30 UTCGHSA enrichmentGHSA-2wcg-j54r-m6wm · GitHub severity HIGH · GitHub CVSS 5.9
  67. 2026-09-18 20:31 UTCEG score recompute
  68. 2026-09-18 19:19 UTCEG score recomputeFirst assessed: 5.9 MEDIUM
  69. 2026-09-18 19:19 UTCMITRE cvelistV5first tracked

Frequently asked(5)

What is CVE-2026-93838?
CVE-2026-93838 is a medium-severity vulnerability; its CVE record was published on September 18, 2026. SGLang versions through 0.5.20 contain an unbounded memory allocation vulnerability in handle_staging_req() that fails to validate chunk_idx from ZMQ STAGING_REQ frames in prefill/decode disaggregation deployments. Attackers with access to the decode engine's internal ZMQ rank port can send a frame…
When was CVE-2026-93838 made public?
CVE-2026-93838 was made public on September 16, 2026: that is the date its CVE record's CNA, VulnCheck, gives for when the vulnerability was made public. The CVE record itself was published on September 18, 2026. EchelonGraph's copy of the record carries a last-modified date of September 24, 2026; the record at its source may have been updated since.
Is CVE-2026-93838 actively exploited?
CVE-2026-93838 is not currently on CISA's Known Exploited Vulnerabilities catalog. FIRST EPSS estimates a 0.7% probability of exploitation in the next 30 days (51st percentile of EPSS-scored CVEs).
What is the CVSS score of CVE-2026-93838?
CVE-2026-93838 has a CVSS base score of 5.9 (the CNA's own assessment, by vulncheck; NVD's own analysis pending).
How do I remediate CVE-2026-93838?
No fix for CVE-2026-93838 is confirmed yet. Until one is published, restrict network exposure of the affected system or apply the vendor's mitigation — for example, keep it off the internet or limit it to trusted networks — and watch the vendor's advisory for the fix. The CVE record's CNA (VulnCheck) lists sgl-project sglang versions up to and including 0.5.20 as affected. No release containing a fix is confirmed on record. The vendor advisories EchelonGraph has for CVE-2026-93838 are linked in the Vendor Advisories panel on this page.

Dependency Blast Radius

Explore the affected products and dependency analysis for CVE-2026-93838

Explore →

Is Your Infrastructure Affected by CVE-2026-93838?

EchelonGraph automatically scans your cloud infrastructure and maps CVE exposure using blast radius analysis.