Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessConnect to cluster as client
-
DeliverySend malicious serialized payload
-
ExploitTrigger improper input validation (CWE-20)
-
ExecutionRead arbitrary member heap/off-heap/JVM memory
-
PersistCorrupt memory or crash cluster member
-
ImpactPossible arbitrary code execution (some EE configs)
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A05:2025 - Injection
ASVS 5.0
V2 - Validation and Business Logic Input validation, allowlisting
|
| Exploitation Conditions | The attacker must be able to establish a client connection to a Hazelcast cluster member (i.e., have network reachability to the member's listening port). In default Community Edition deployments where Hazelcast Security and client authorization are not enabled, any reachable client can connect without authentication; where client authorization is enforced, a low-privileged client credential is required. The arbitrary-read and crash (DoS) impacts affect both slim and full distributions of Community and Enterprise Editions; the memory-corruption-to-RCE impact is limited to 'some Hazelcast Enterprise Edition configurations.' The patch introduces/strengthens a zero-config Compact serialization class filter (zero-config-filter allowlist/blacklist), indicating that unrestricted zero-config Compact serialization is a relevant exposure. Limiting factors that reduce real-world risk: firewalling member ports from untrusted networks, enforcing client authorization, and defining an explicit zero-config Compact serialization allowlist - all recommended in the vendor hardening guide. |
| Impact | This is a genuine high-priority vulnerability, not an inflated CVSS score. The vendor-supplied CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H, score 9.3 Critical) describes a network-reachable, low-complexity flaw with high confidentiality, integrity, and availability impact and no user interaction. The impact is well-substantiated by the GHSA advisory: arbitrary reads of a member's Java heap, off-heap data, and JVM process address space (confidentiality), member crashes (availability/DoS), and, in some Enterprise Edition configurations, memory corruption leading to possible RCE (integrity). A notable signal conflict exists: the official vector states PR:N (unauthenticated), yet both the NVD description and GHSA advisory repeatedly characterize the attacker as a 'low-privileged malicious client able to connect to a cluster' - implying at least the ability to establish a client connection, which in hardened deployments requires authorization. This should be verified against your deployment: in default Community Edition (no security enabled) PR:N is realistic, but where Hazelcast Security and client authorization are enforced, the practical barrier is higher (PR:L). No EPSS score, no CISA KEV listing, and no public exploit (POC) were provided, so there is no evidence of active exploitation - the urgency is driven by impact severity and patch availability rather than observed in-the-wild activity. Net: treat as a priority patch for any cluster whose member ports are reachable by untrusted or semi-trusted clients. |
| Exploit ScenarioDerived | A malicious or compromised client that can reach a Hazelcast cluster member's listening port connects and sends crafted, improperly-validated serialized data. The server-side flaw lets the attacker read arbitrary regions of the member's heap, off-heap store, and JVM address space (exfiltrating secrets and other tenants' data) and crash members for denial of service; in some Enterprise Edition configurations it can corrupt memory for possible remote code execution. No public exploit code was identified at time of analysis. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Oct 08, 2026
Arbitrary memory disclosure and denial of service in Hazelcast allow a client that can reach a cluster member's listening port to read arbitrary data from that member's Java heap, off-heap storage and JVM process address space, to crash members, and - in some Hazelcast Enterprise Edition configurations - to corrupt memory with possible remote code execution. All editions prior to 5.4.5, 5.5.10, 5.6.1 and 5.7.0 are affected in both the slim and full distributions, and the assessed CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H) reflects a low-privileged client credential where client authorization is enforced, while in default deployments with Hazelcast Security disabled any reachable client can connect without credentials; the vendor's CVSS 4.0 score of 9.3 assumes PR:N. There is no public exploit identified at time of analysis and the issue is not listed in CISA KEV, but vendor-released patches exist (5.4.5, 5.5.10, 5.6.1, 5.7.0) and the vendor's own hardening guidance - enabling security, enforcing client authorization and defining an explicit zero-config Compact serialization allowlist - directly addresses the exposure, making this a genuine high-priority vulnerability rather than an inflated score.
RemediationAI
Vendor-released patch: upgrade to Hazelcast 5.4.5, 5.5.10 or 5.6.1 on the corresponding maintenance line, or to 5.7.0 (Community Edition users should move to 5.7.0, since that is the release the GHSA names for CE); the fix is carried in commit 361979da12f18950c24719db832ca6c5e7c0534f and described at https://github.com/hazelcast/hazelcast/security/advisories/GHSA-6v25-8wq6-xq4j, with extended-support customers directed to Hazelcast Support for older-version patches. If upgrading cannot be done immediately, the vendor's hardening guide (https://docs.hazelcast.com/hazelcast/5.7/secure-cluster/hardening-recommendations) gives concrete compensating controls: enable Hazelcast Security and enforce client authorization so that only authenticated, authorized clients can connect (trade-off: all clients need credentials configured and unauthenticated or legacy clients will be rejected until updated); define an explicit allowlist for zero-config Compact serialization using the zero-config-filter with defaults-disabled plus class/package/prefix allow- and deny-lists (trade-off: any legitimate class not on the allowlist will fail to deserialize, so the list must be derived from actual application traffic); firewall the member listening ports so only trusted client and member subnets can reach them and refrain from exposing clients on internet-facing or non-secure networks (trade-off: remote or partner clients, WAN replication and cross-DC topologies that rely on open member ports will break until rules are widened); and disable any Hazelcast features that are not in use to shrink the attack surface. Note that until a compensating control is in place, any host with network reachability to a member port - and in default deployments without security enabled, without any credential at all - should be treated as able to trigger arbitrary memory reads or member crashes.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 2 packages in maven, euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
Hazelcast is a Java-based in-memory data grid / real-time data platform in which cluster members exchange serialized objects with clients over the Hazelcast client protocol. The flaw is classified as CWE-20 (Improper Input Validation): the member does not adequately validate data supplied by a connecting client before using it, so a malicious client can steer member-side operations into reading memory outside the intended data structures, exposing Java heap contents, off-heap (native) buffers and other address space of the JVM process. The same malformed input can trigger member crashes, and in certain Hazelcast Enterprise Edition configurations the corruption can extend into memory contents with the possibility of arbitrary code execution, hence the RCE, Java and Hazelcast tags on this record. The upstream fix commit 361979da12f18950c24719db832ca6c5e7c0534f centers on the serialization configuration layer, adding and wiring up a zero-config Compact serialization filter (JavaSerializationFilterConfig, ClassFilter with class/package/prefix allow- and deny-lists, and a zero-config-filter element with defaults-disabled), which indicates that unrestricted zero-config Compact serialization of client-supplied classes was the relevant exposure path. The CPE cpe:2.3:a:hazelcast:hazelcast:*:*:*:*:*:*:*:* is recorded with wildcard versioning, so exact affected ranges must be taken from the Maven coordinates and advisory text: com.hazelcast:hazelcast is affected at 5.6.0, from 5.5.0 up to (but excluding) 5.5.10, and below 5.4.5, with 5.7.0 listed as the fixed release in the package metadata.
Know when Hazelcast gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
Network-reachable, low-complexity flaw; PR:L reflects the advisory's 'low-privileged client' (not PR:N) since a client connection is required; full heap read (C:H), memory corruption/RCE (I:H), and member crash (A:H).
Primary rating from Vendor (GitHub_M).
CVSS VectorVendor: GitHub_M
Lifecycle Timeline
2DescriptionCVE.org
Hazelcast is a unified real-time data platform combining stream processing with a fast data store. Prior to 5.4.5, 5.5.10, and 5.6.1, improper validation of data supplied by a malicious client able to connect to a cluster allows arbitrary reads from a cluster member's Java heap, off-heap data, and JVM process address space. The same flaw can crash cluster members and, in some Hazelcast Enterprise Edition configurations, corrupt memory with possible arbitrary code execution. Both slim and full distributions are affected. This issue is fixed in versions 5.4.5, 5.5.10, 5.6.1, and 5.7.0.
Blast Radius
ecosystem impact- 236 maven packages depend on com.hazelcast:hazelcast (81 direct, 155 indirect)
Ecosystem-wide dependent count for version 5.6.0.
Articles & Coverage 1
Technical POC Analysis
The advisory classifies the flaw as CWE-20 Improper Input Validation.
Full technical POC analysis - root cause, step-by-step exploitation and patch breakdown - is part of the Pro and Business plans.
Executive ViewAI
Hazelcast, the in-memory data grid that underpins several of our customer-facing and internal applications, carries a critical flaw (CVE-2026-107726, CVSS 9.3) that allows a connected client to read arbitrary memory from a cluster member and to crash or, in some Enterprise Edition configurations, corrupt that member with possible remote code execution. Every deployment below versions 5.4.5, 5.5.10, 5.6.1 or 5.7.0 is exposed in both the slim and full distributions, and because Hazelcast Security is off by default, a reachable client may require no credentials at all, while where authorization is enforced the CVSS vector (PR:L) indicates only a low-privileged credential is needed. Leadership should care because disclosed heap and off-heap contents can expose credentials, cryptographic keys and regulated customer data, while a crashed or compromised member can take down the services that depend on the cluster.
Apply vendor patch.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Remote Code Execution represents the critical moment when an attacker successfully runs arbitrary code on a target system without physical access.
Improper Restriction of XML External Entity Reference in GitHub repository hazelcast/hazelcast in 5.1-BETA-1. Rated crit
Authorization bypass in the Hazelcast IMap Predicates API allows a client holding only limited cluster credentials to ex
Hazelcast Management Center through 6.0 allows remote code execution via a JndiLoginModule user.provider.url in a hazelc
The LDAP authentication method in LdapLoginModule in Hazelcast IMDG Enterprise 4.x before 4.0.3, and Jet Enterprise 4.x
The Connection handler in Hazelcast and Hazelcast Jet allows a remote unauthenticated attacker to access and manipulate
In Hazelcast through 5.0.4, 5.1 through 5.1.6, and 5.2 through 5.2.3, executor services don't check client permissions p
In Hazelcast through 5.0.4, 5.1 through 5.1.6, and 5.2 through 5.2.3, configuration routines don't mask passwords in the
EUVD-2026-95314
GHSA-6v25-8wq6-xq4j
NVD
CVE.org
EPSS
Vulnrichment
VulDB
Exploit-DB