Security readout for executives and security teams
Plain-English summary
SGLang has a critical flaw in an optional expert-parallel backup feature. If enabled and reachable over the network, an unauthenticated attacker could cause remote code execution through unsafe pickle handling. The source bundle does not identify a patch, fixed version, or active exploitation.
Executive priority
Prioritize immediately for any internet-facing or broadly reachable SGLang environment. Business risk is unauthorized code execution, with high confidentiality and integrity impact. Lower priority only if the feature is disabled or isolated from untrusted networks.
Technical view
The issue is CWE-502 deserialization in SGLang's expert backup manager. A ZeroMQ PULL socket is exposed on a routable interface without authentication or deserialization safeguards, enabling unauthenticated network RCE when the feature is enabled and reachable. CVSS is 9.1: AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N.
Likely exposure
Exposure appears limited to SGLang deployments where the expert-parallel backup subsystem is enabled and network reachable. The affected version data only lists "0", so precise version ranges are unclear from the supplied sources.
Exploitation context
The bundle marks KEV as false and provides no cited evidence of active exploitation. The described attack path is low-complexity and unauthenticated, so reachable deployments should be treated as high-priority until vendor guidance is confirmed.
Researcher notes
Evidence is strong for the vulnerability class and attack preconditions, but incomplete for affected versions and remediation. Avoid assuming all SGLang deployments are affected; validate feature enablement, binding behavior, and reachability.
Mitigation direction
- Disable the expert-parallel backup subsystem if it is not required.
- Restrict network reachability to trusted hosts only.
- Block unauthenticated access to the exposed ZeroMQ listener.
- Check SGLang and CERT guidance for patches or configuration changes.
- Treat exposed instances as urgent until remediation is verified.
Validation and detection
- Inventory SGLang deployments and identify expert-parallel backup usage.
- Confirm whether the backup listener binds to a routable interface.
- Review firewall rules and service exposure paths.
- Check SGLang version and compare against vendor guidance.
- Verify logs for unexpected backup subsystem connections.
Public sources used
Generated from the cited source records. This long-tail analysis has not been individually reviewed by a named human.
Conservative CVE-to-ATT&CK context
These mappings and lookup hints may be relevant to the vulnerability behavior, CWE, affected product, or exposure path. Glexia-inferred context is not an official MITRE, ATT&CK, CWE, or CVE Program mapping.
ATT&CK lookup starting points
Use these exact CWE pages and searches to review the Glexia ATT&CK library from this CVE's weakness and description context.
CWE-502: Code execution behavior lookup
Code execution and unsafe deserialization weaknesses often justify reviewing execution behavior and process telemetry. Open the exact CWE lookup page first, then review the ATT&CK searches from that MITRE weakness context. This is a Glexia lookup hint, not an official ATT&CK mapping.
Open ATT&CK lookupExecution behavior lookup
The CVE wording references code or command execution, so execution technique review may help defensive triage. This is a Glexia inferred lookup path, not an official MITRE, ATT&CK, or CVE Program mapping.
Open ATT&CK lookupCVE-2026-14890 mapping review
Open the CVE-to-ATT&CK bridge for reviewed, inferred, or future official mappings tied to this CVE.
Open ATT&CK lookup- Severity
- Critical
- CVSS
- 9.1 (3.1)
- Known Exploited
- No
- Published
Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N
CNA and ADP enrichment extracted from CVE v5
These fields come from the CVE record and ADP containers, not from Glexia's Take. They preserve time-varying source decisions such as CISA SSVC, KEV status, CVSS metrics, and provider references.
SSVC decision data
CVSS vector scores
1 official scoreWe collect every scored CVSS vector available in the official CNA and ADP containers. When more than one version is present, the table keeps the source vectors side by side instead of collapsing them into the highest score.
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N3.95.2CISA-ADPVulnerability scoring details
Base CVSS 3.1 score
9.1CriticalVector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N
Vulnerability timeline
Timeline events are normalized from CVE metadata, CNA source timelines, ADP timelines, and KEV metadata when present.
- CVE reservedCVE Program
The CVE ID was reserved by the assigning CNA.
- CVE publishedCVE Program
The CVE record was published.
- CVE updatedCVE Program
The CVE record metadata indicates this as the latest update time.
ADP provider summaries
Source materials
- CVE List V5 sourceCVE List V5
- https://github.com/sgl-project/sglang/blob/main/python/sglang/srt/elastic_ep/expert_backup_manager.pyCVE reference, certcc
- https://www.kb.cert.org/vuls/id/326070CVE reference, CVE
Products and packages named in the record
CWE details
CWE links open Glexia weakness intelligence pages with official CWE context, developer remediation guidance, and related CVE mappings.
Deserialization of Untrusted Data
Deserialization of Untrusted Data represents a recurring weakness pattern that can create exploitable paths when design, validation, or implementation controls are missing.