Reflected XSS in Netron desktop application versions ≤9.1.2 executes attacker-controlled JavaScript when a victim opens a crafted neural network model file containing malicious node names, type labels, or value descriptions. Because Netron runs on Electron (a Chromium-based runtime), the injected script operates within a privileged renderer process capable of port-scanning local networks and, with a separately-acquired Chrome n-day, achieving full remote code execution on the host machine. HiddenLayer reported the vulnerability and a vendor fix is available via GitHub commit cd14bad8c91; no public standalone exploit has been confirmed and this CVE does not appear in CISA KEV.
TLS certificate validation is absent by default in Spring AMQP's Log4j2 appender when shipping log events to a RabbitMQ broker, leaving all log traffic fully exposed to man-in-the-middle interception. Spring AMQP versions 2.4.18 and earlier, 3.2.0-3.2.12, 4.0.0-4.0.4, and 4.1.0 are affected when the Log4j2 appender is used with its documented default configuration. No public exploit has been identified at time of analysis, but the high confidentiality and integrity CVSS impact reflects that intercepted log streams may contain API keys, session tokens, and other sensitive application data.
Zip extraction in AWS diagram-as-code (awsdac) versions 0.10 through 0.23 is vulnerable to relative path traversal (CWE-23), enabling a third party to write arbitrary files outside the intended extraction directory by embedding crafted path sequences in zip entry names within a malicious diagram bundle. Exploitation requires user interaction - a victim must actively process the attacker-supplied bundle - making this a social-engineering-dependent local attack. AWS has released a patched version 0.24; no active exploitation or public exploit code has been identified at time of analysis.
Out-of-bounds write in the SmiFlash SMM (System Management Mode) module affects ASUS FA507NU and FA507NV laptop BIOS firmware across all versions. A local administrator can dispatch a crafted Software SMI (SW SMI) request with an oversized length value, causing the SmiFlash handler to write beyond its allocated SMRAM buffer boundary, resulting in a system crash (BSOD) or irreversible BIOS flash corruption. The affected models are designated 'unsupported-when-assigned,' meaning these products were already outside active vendor support when this CVE was issued; no public exploit has been identified at time of analysis.
Sensitive KMS provider credentials supplied for client-side field-level encryption (CSFLE) in the MongoDB C# Driver are reproduced in plaintext within the driver's diagnostic string representation of its client settings, rather than being masked like other secret fields. Any party with read access to application logs, diagnostic output, or process memory dumps can recover these credentials and independently invoke the key management service to unwrap data-encryption keys, compromising all data protected by those keys. No public exploit code or active exploitation in CISA KEV has been identified at time of analysis; exploitation is bounded by the requirement that CSFLE be configured and that the attacker have access to diagnostic output.
Spring Integration's jCIFS-based SMB client silently negotiates down to legacy SMB1/CIFS when applications omit an explicit `smbMinVersion` setting, exposing file transfer sessions to NTLM relay attacks and in-transit content tampering by a network-positioned adversary. Affected versions span Spring Integration 6.4.0-6.4.12, 6.5.0-6.5.10, 7.0.0-7.0.5, and the standalone release 7.1.0 - covering all currently active Spring Integration release trains at time of disclosure. No CISA KEV listing or public exploit code specific to this CVE has been identified at time of analysis, though NTLM relay is a mature, extensively tooled attack class; EPSS data was not available in the sourced intelligence.
Arbitrary code execution across entire WordPress multisite networks is possible in the Defender Security plugin (versions 5.0.0 through 6.1.x) due to a missing access control check on a network-wide setting. Any site-level administrator on a multisite installation - a role with significantly lower trust than a network/super administrator - can modify this setting to inject and execute code affecting all sites on the network. No public exploit has been identified and EPSS stands at 0.17% (6th percentile), but SSVC rates technical impact as 'total', consistent with the network-wide RCE consequence if exploited by a malicious or compromised site admin.
Cleartext cluster key disclosure in Wazuh 4.14.0-4.14.6 allows any low-privilege API user with the default readonly or cluster_readonly role to retrieve the shared secret that authenticates and encrypts all inter-node cluster traffic. The GET /cluster/local/config endpoint was missing the mask_sensitive_config decorator applied to every sibling config endpoint, causing the cluster.key field to be returned unredacted to accounts that are explicitly denied update-config permission. The severity of this information disclosure is amplified by prior Wazuh advisories that establish cluster-peer remote code execution chains contingent on possessing the cluster key, making this a practical stepping stone to full cluster compromise. No public exploit has been identified at time of analysis and the vulnerability is not in the CISA KEV catalog.
Denial-of-service in Spring for Apache Kafka's `DeadLetterPublishingRecovererFactory` exposes consumer JVMs across all supported branches (2.8.x through 4.1.0) to crash or hang when processing messages with a crafted `retry_topic-original-timestamp` header. An authenticated Kafka producer can inject a ConsumerRecord carrying an oversized or malformed header byte array, which is passed without any length or format validation directly to Java's `BigInteger(byte[])` constructor, triggering heap memory exhaustion or an unhandled exception that disrupts message consumption. No public exploit code has been identified and the CVE does not appear in CISA KEV at time of analysis, but low attack complexity combined with broad version coverage makes this a credible availability threat for any Spring Kafka deployment using retry or dead-letter queue features.
Arbitrary file write in the libreoffice-convert npm package (all versions prior to 1.8.2) allows any caller who supplies a user-controlled `fileName` option to escape the intended temporary directory and overwrite arbitrary filesystem paths accessible to the Node.js process. The root cause is unsanitized use of `path.join(tempDir.name, fileName)` - a `../`-laden filename resolves outside the temp dir without restriction. With no public exploit identified at time of analysis and no CISA KEV listing, risk is contingent on whether consuming applications expose `fileName` to untrusted input, but the write primitive is trivially exploitable in document-conversion web services.
Cleartext credential exposure in Spring AMQP affects applications that use the RabbitMQ management aliveness check: when that check fails at runtime, the configured admin password is embedded verbatim in the thrown Java exception message, making it visible to anyone with access to exception output. The vulnerability spans multiple major release lines - 2.4.x, 3.2.x, and 4.0.x through 4.1.0 - and was reported by security@vmware.com, the Spring Security team. No public exploit code has been identified at time of analysis and CISA KEV status is unconfirmed, but a successful credential leak would yield full RabbitMQ administrative control.
Link credit exhaustion in Spring AMQP 4.1.0 silently stalls AMQP listener containers when a container-level ErrorHandler is active. Each message delivery whose processing throws an exception permanently consumes one link credit without returning it to the broker; once the pool reaches zero (after the default 100 credits), the broker halts delivery while isRunning() continues to report true, masking the outage from standard health checks. No active exploitation has been confirmed (no CISA KEV listing, no public POC), but this is a realistic denial-of-service condition for any Spring-based messaging consumer running the affected version with an ErrorHandler configured.
Path traversal in Spring AI's ResourceCacheService allows an authenticated low-privileged network attacker to write files outside the intended cache directory. The getCacheName() method appends URI fragments verbatim - including '../' sequences and path separators - to a base directory path before passing the result to new File(resourceParentFolder, newFileName) and writing downloaded bytes there. Affected versions span Spring AI 1.0.9 and earlier, 1.1.0 through 1.1.8, and 2.0.0; no public exploit or CISA KEV listing has been identified at time of analysis.
Spring Integration's UnZipTransformer component fails to impose any limit on decompressed entry size or entry count, enabling a low-privileged remote attacker to trigger JVM heap exhaustion via a crafted zip bomb archive. Affected versions span 6.4.0-6.4.12, 6.5.0-6.5.10, and 7.0.0-7.1.0 across the widely-used Spring enterprise Java framework. The attack primitive is well-understood and trivially constructable, though exploitation requires access to an integration endpoint that routes payloads through UnZipTransformer; no public exploit code has been identified at time of analysis and this CVE is not listed in the CISA KEV catalog.
SQL injection in the NL2SQLTool component of crewai-tools v1.10.2rc1 exposes any application built on this AI agent framework to arbitrary database query execution by remote unauthenticated attackers. The unsanitized sql_query argument is passed directly to the underlying database engine, enabling classic injection attacks against the connected data store. No public exploit code has been identified at time of analysis, and the EPSS score of 0.23% (13th percentile) indicates low current exploitation activity, though the zero-authentication requirement elevates exposure for internet-facing deployments.
Authorization bypass via user-controlled key in Summit Security Systems AdisyonPro (all versions before v5.21.0) allows unauthenticated remote attackers to access application functionality that should be restricted by access control lists, resulting in integrity and availability impact. The flaw is classified as CWE-639, an Insecure Direct Object Reference (IDOR)-class weakness in which the application accepts user-supplied keys to gate access to protected resources without adequate server-side authorization validation. No public exploit code or active exploitation has been identified at time of analysis.
Foreman's template revision endpoint, as shipped in Red Hat Satellite 6, exposes historical template contents across organizational and location boundaries due to missing object-level authorization (IDOR/BOLA, CWE-639). An authenticated user holding any template-related permission - as minimal as view_ptables - can retrieve audit-history revisions owned by other organizations simply by supplying a guessable audit ID. Template revisions may contain credentials, secrets, or sensitive configuration data, making the confidentiality impact high despite the low attack complexity. No public exploit has been identified at time of analysis, though the low-privilege entry bar lowers the practical exploitation threshold significantly.
Stored Cross-Site Scripting in Fluent Boards Pro (WordPress plugin by WP Manage Ninja) versions up to and including 2.0.11 allows a subscriber-level authenticated attacker to inject persistent malicious scripts into board content. Because the scope changes (S:C in the CVSS vector), the injected payload executes in the browser context of other users - most critically administrators - who subsequently view the compromised content, enabling session hijacking or privilege escalation. No public exploit code or active exploitation has been identified at time of analysis, but the low privilege requirement (subscriber account) broadens the realistic attacker population on multi-user WordPress sites.
Insufficient data exists to characterize CVE-2026-19953 beyond its association with an Ubuntu vendor report. No description, CVSS vector, CWE, or reference URLs are available in the provided intelligence. The affected component, impact class, and exploitation conditions are entirely unknown at this time.
Path traversal in Veno File Manager 4.4.9 exposes arbitrary server files to authenticated super administrators via two crafted HTTP requests targeting /vfm-admin/index.php and /vfm-admin/ajax/streamvid.php. The vulnerability allows directory escape through unsanitized path parameters, enabling disclosure of sensitive files such as configuration files, credentials, or system data accessible to the web server process. A public proof-of-concept repository exists on GitHub (jfs-jfs/CVE-2026-37066), and no public exploit has been confirmed as actively exploited per CISA KEV.
Header type injection in Spring for Apache Kafka's default mapper configuration allows any Kafka producer with access to a consumed topic to inject java.net.InetAddress objects through the spring_json_header_types message header, triggering DNS resolution on the consumer host. All applications using @KafkaListener with the default JsonKafkaHeaderMapper or DefaultKafkaHeaderMapper - spanning versions 2.8.x through 4.1.0 - are affected without any non-default configuration required. No public exploit code has been identified at time of analysis, and the CVE is not listed in CISA KEV.
Unauthorized file access in Veno File Manager 4.4.9 exposes any user-uploaded file to other authenticated users via the /vfm-admin/ajax/streamvid.php streaming endpoint, which applies no per-user authorization enforcement before returning file content. An attacker holding a valid account on the same instance can craft a GET request to retrieve files belonging to any other user, provided the target file's path and filename are known. A public proof-of-concept repository exists at github.com/jfs-jfs/CVE-2026-37070; no active exploitation is confirmed by CISA KEV.
Integer overflow in FFmpeg's libavfilter/vf_scale.c video scaling filter allows a remote attacker to crash the FFmpeg process by delivering a crafted video file that a user then opens or processes. Affected is FFmpeg development build N-122528-gdd2976b9e1, a git-snapshot identifier rather than a stable release. No public exploit and no active exploitation are identified at time of analysis; EPSS sits at 0.15% (4th percentile), confirming low observed exploitation probability despite the ubiquitous deployment footprint of FFmpeg.
Division-by-zero in FFmpeg's libswscale library allows remote attackers to crash the process by supplying a crafted media file that triggers an unhandled zero-divisor during scaling context initialization. The fault occurs in ff_sws_init_single_context within /libswscale/utils.c when malformed input parameters produce a zero-valued denominator, resulting in a process crash and denial of service. No active exploitation has been confirmed, and EPSS at 0.15% (4th percentile) indicates negligible observed exploitation activity.
Session tokens in HCL IntelliOps Event Management (IEM) versions 1.3 and 1.4 persist as valid on the server after a user logs out or an account is deleted, allowing an attacker who has previously captured a legitimate session token to replay it and maintain unauthorized access beyond the intended session lifecycle. The root cause is CWE-613 (Insufficient Session Expiration), where server-side session records are not destroyed upon logout or account removal. No public exploit code or active exploitation has been identified at time of analysis, and the CVSS 6.4 Medium score reflects meaningful constraints on exploitation including high attack complexity and required user interaction.
WatchGuard Dimension's web login endpoint permits unlimited automated password guessing against user accounts because rate-limiting and account lockout are both absent in the default product configuration. All versions of Dimension (CPE wildcard, no version boundary) are affected in default-configured deployments. An unauthenticated remote attacker can systematically guess credentials and gain unauthorized account access; the CWE-203 classification further indicates the endpoint may expose observable discrepancies that enable username enumeration, reducing the effort required for targeted credential attacks. No public exploit has been identified at time of analysis and no CISA KEV listing exists.
Regular expression denial of service in NLTK's Pl196xCorpusReader allows unauthenticated remote attackers to exhaust CPU resources by supplying malformed TEI corpus data containing many unmatched opening tags. Applications exposing NLTK's words() or tagged_words() public APIs over a network boundary-such as NLP services processing user-supplied corpora-are at risk of near-quadratic runtime growth in the read_block method. No public exploit code or CISA KEV listing has been identified at time of analysis; patch 3.10.3 is available.
Quadratic CPU exhaustion in NLTK before 3.10.3 allows denial of service via malformed XML corpus files processed by XMLCorpusView._read_xml_fragment(). The method rescans the entire accumulated XML buffer on every 1 KiB block read, creating O(n²) time complexity that scales severely with input size. Applications exposing XML corpus ingestion to untrusted input - such as those using BNCCorpusReader - are at realistic risk; no public exploit has been identified and this is not listed in CISA KEV.
Unsafe header deserialization in Spring Integration's EmbeddedHeadersJsonMessageMapper exposes applications to header injection across all active release trains (5.5.x through 7.1.0). The decodeNativeFormat method constructs GenericMessage objects with MutableMessageHeaders populated directly from attacker-controlled JSON-deserialized header maps, bypassing sanitization by default in the constructor. Low-privileged network attackers who can submit crafted byte payloads to an affected consumer channel can manipulate message headers, yielding partial confidentiality, integrity, and availability impacts. No public exploit has been identified at time of analysis and this CVE is not listed in CISA KEV.
Path traversal in arben-adm mcp-sequential-thinking versions up to and including 0.5.0 allows remote low-privileged attackers to read or write arbitrary files on the server by supplying unsanitized path sequences to the `import_session` and `export_session` functions in `server.py`. A publicly disclosed proof-of-concept exists (GitHub issue #17), and the fix is available in v0.6.0 via commit 2fad3ee8. The CVE is not listed in CISA KEV, indicating no confirmed mass exploitation, but the public PoC meaningfully elevates practical risk above the moderate CVSS score alone.
Path traversal in RooCodeInc Roo-Code through version 3.51.1 allows remote unauthenticated attackers to read and write files outside the intended working directory by manipulating input processed by the path.resolve function in ApplyPatchTool.ts. The vendor has confirmed they archived the repository and discontinued all support, meaning no patch will ever be released. Publicly available exploit code exists via a GitHub Gist, and no vendor-released patch identified at time of analysis - organizations still running this tool are permanently exposed.
Code injection in RooCodeInc Roo-Code through version 3.51.1 allows remote attackers to execute arbitrary code via the ExecaTerminalProcess function in the README File Handler, triggered by passive user interaction with a crafted README file. The vendor has explicitly archived the repository and confirmed no patch will be issued, leaving all installed instances permanently unpatched. A public proof-of-concept exploit is available on GitHub; however, no active exploitation has been confirmed by CISA KEV, and the CVSS 4.0 score of 2.1 reflects limited impact ratings across all three CIA pillars.
Open redirect in Spring Authorization Server 1.5.0-1.5.7 allows remote unauthenticated attackers to redirect victims to attacker-controlled sites by crafting authorization requests that pair an invalid request_uri with an unvalidated redirect_uri. The CVSS scope change (S:C) reflects that the redirect escapes the authorization server's security boundary, enabling high-trust phishing attacks because victims see the legitimate authorization server domain before being redirected. No public exploit has been identified at time of analysis, and exploitation requires user interaction.
Reflected cross-site scripting (XSS) in Spring Security's built-in OAuth2 Authorization Server consent page allows an unauthenticated remote attacker to execute arbitrary JavaScript in a victim's browser by delivering a crafted authorization request URL. Affected versions include Spring Security 7.1.0 and 7.0.0 through 7.0.6; the consent page reflects user-controlled values (such as client name or scope parameters) directly into the HTML response without encoding. No public exploit has been identified at time of analysis, and this CVE does not appear in the CISA KEV catalog, but the changed-scope CVSS vector reflects the elevated risk of token and session theft in OAuth2 flows.
Reflected XSS in Seres Software syWEB through version 27082026 allows unauthenticated remote attackers to inject and execute arbitrary JavaScript in victims' browsers by inducing them to click a crafted URL. The vulnerability is compounded by the vendor's confirmation that syWEB is end-of-life and unsupported, meaning no patch will ever be released. No public exploit code has been identified at time of analysis, and the product has no CISA KEV listing.
Cross-site scripting (Cross Zone Scripting variant) in BilPark Informatics Technologies' DoXBASE allows unauthenticated remote attackers to inject and execute malicious scripts in a victim's browser, affecting all versions through 27082026. The CVSS 3.1 score of 6.1 reflects network-based exploitation requiring victim interaction, with a changed scope indicating the injected script executes across browser security zones. No patch is available and the vendor did not respond to TR-CERT's disclosure, leaving organizations with no official remediation path.
Reflected XSS in Spring Framework's MVC and WebFlux data-binding error rendering exposes users of vulnerable applications to arbitrary HTML and JavaScript injection. All active Spring Framework release trains are affected - versions 5.2.x through 7.0.8 - making the exposure surface extremely broad across enterprise Java deployments. An unauthenticated attacker (per CVSS PR:N) can craft a malicious URL containing injected script payloads that, when a victim clicks it, execute in the victim's browser context via reflected field-error output. No public exploit code has been identified and EPSS sits at 0.17% (7th percentile), indicating limited observed exploitation activity at time of analysis.
Open redirect in Spring Framework's UrlHandlerFilter allows unauthenticated remote attackers to silently redirect authenticated users to attacker-controlled URLs. Affected are Spring Framework 7.0.0-7.0.8 and 6.2.0-6.2.19 in both Spring MVC and Spring WebFlux deployments where UrlHandlerFilter is configured with broadly matching URL patterns. No public exploit code exists and SSVC indicates no known active exploitation; however, the no-privilege, network-reachable attack vector makes this a credible phishing enabler in affected applications. EPSS sits at 0.15% (5th percentile), consistent with low-automated-exploitation likelihood.
Open redirect in Spring Framework's UrlFileNameViewController exposes users of affected Spring MVC applications to phishing and credential-harvesting attacks across all active Spring Framework release trains (5.2.x through 7.0.x). The flaw allows an unauthenticated attacker to craft a URL that causes the vulnerable application to redirect victims to an arbitrary external domain, exploiting the controller's path-to-view-name resolution when end-of-path mapping is used without a configured prefix. No active exploitation has been confirmed - EPSS sits at 0.14% (4th percentile) and CISA SSVC reports exploitation status as none, making this a targeted patching priority rather than an emergency.
TLS bypass in Grafana's ClickHouse Datasource plugin (versions 3.1.0 through 4.20.0) silently transmits database traffic in cleartext when the Native protocol is routed through a PDC or secure SOCKS proxy, despite TLS being configured. The underlying connection library ignores the TLS request without raising an error, exposing ClickHouse usernames, passwords, SQL queries, and query results to any attacker positioned on the network hop between the proxy and the database server. No public exploit has been identified at time of analysis and the vulnerability is not listed in the CISA KEV catalog.
Path traversal in Eclipse SW360's optional filesystem attachment store allows authenticated low-privilege users to write files to arbitrary paths on the host operating system by crafting malicious filenames during attachment upload. Affected versions span 19.0.0 through 20.1.0 exclusively when the non-default `enable.attachment.store.to.file.system` configuration key is active. The CVSS 4.0 score of 6.0 (AT:P, PR:L, VI:H) reflects meaningful integrity impact gated behind a specific configuration requirement; no public exploit has been identified at time of analysis.
Path traversal in MAA-AI MaaMCP (versions up to 1.1.1.dev6+g2e4a41287) allows low-privileged remote users to read or write files outside the intended pipeline directory by supplying crafted path inputs to the save_pipeline and load_pipeline functions in pipeline_tools.py. Publicly available exploit code exists (referenced in GitHub issue #31), though no active exploitation has been confirmed by CISA KEV. The CVSS 4.0 score of 5.1 (Medium) reflects the authentication requirement and limited per-impact severity, but the public POC elevates practical risk above the base score alone.
Code injection in RooCodeInc Roo-Code up to version 3.51.1 allows a remote low-privileged attacker to execute injected code via the `fetch_instructions` function within the MCP (Model Context Protocol) Integration Trust Model component (`malicious_mcp_server.py`). A publicly available proof-of-concept exploit exists on GitHub. Critically, the vendor has explicitly confirmed the product is no longer maintained - the repository is archived - meaning no patch will ever be released, permanently leaving all deployed instances vulnerable.
Code injection in RooCodeInc Roo-Code up to version 3.51.1 allows a remote, low-privileged attacker with passive user interaction to inject and execute arbitrary code via the `optimizeQuery` function in the CodeIndexManager component (`src/utils/helpers.ts`). A public proof-of-concept exploit exists on GitHub. Critically, the vendor has explicitly archived the repository and will not issue a patch, leaving all users of this product permanently unpatched.
Denial of service in Silicon Labs SiSDK v2026.6.0 and earlier affects EFR32MG24 and EFR32MG26 wireless SoCs running concurrent multiprotocol Zigbee and Thread. Under high network traffic loads, ACK packets are dropped, causing the device networking stack to fail and rendering the IoT endpoint unresponsive. No public exploit has been identified and active exploitation has not been confirmed by CISA KEV, but the patch is available from the vendor.
OS command injection in AlexGladkov claude-in-mobile 3.10.2 allows a local low-privileged attacker to execute arbitrary shell commands by supplying metacharacter-laced parameters to the tool's ADB and iOS simulator CLI subsystem. The Rust CLI modules (android.rs, ios.rs) interpolated caller-supplied strings - package names, permission identifiers, XML filenames, preference keys, SQLite database paths, and relative file paths - directly into device-side sh -c invocations without sanitization. A public proof-of-concept exists (GitHub issue #40), and a vendor-released patch (v3.10.3) is available.
EAZ EazyFix 12.9 exposes a security feature bypass (CWE-325: Missing Required Cryptographic Step) that allows a highly privileged local attacker to disable Secure Boot, undermining the firmware-level integrity chain that prevents unauthorized boot components from executing. The MSRC entry at https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-25250 confirms Microsoft ecosystem relevance, suggesting the flaw interacts with Windows Secure Boot validation. No public exploit is identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog.
Timing side-channel exposure in Spring Security allows remote unauthenticated attackers to progressively infer security-sensitive string values - such as CSRF tokens, bearer tokens, or other secrets compared at the application layer - by statistically measuring the time differential between rejection responses. Multiple active Spring Security release trains are affected, spanning versions 5.7.x through 7.1.x, representing a broad swath of production Spring-based Java applications. No public exploit code has been identified at time of analysis, and CISA KEV status is absent; however, the confidentiality impact is rated High in the CVSS vector, reflecting that successful exploitation could fully compromise protected secrets under favorable network conditions.
Denial of service in jwcrypto's JWE.deserialize() function allows a remote unauthenticated attacker to crash any service that processes untrusted JWE tokens by submitting a malformed compact serialization containing an excessive number of period delimiters, triggering uncontrolled memory allocation and a MemoryError. Affected deployments include Red Hat Ansible Automation Platform 2, RHEL 9 and 10, OpenShift AI (RHOAI), and OpenStack Platform 16.2 as disclosed via Red Hat Bugzilla 2524147 and upstream advisory GHSA-96rv-c4vc-h4f4. No public exploit code or CISA KEV listing exists at time of analysis, and impact is confined to availability with no confidentiality or integrity exposure.
Resource exhaustion in Spring for GraphQL's WebSocket client allows unauthenticated remote attackers to trigger a Denial of Service when the keepAlive feature is enabled. Three release trains are affected: 1.3.x through 1.3.9, 1.4.x through 1.4.6, and 2.0.x through 2.0.4, as confirmed by the VMware/Spring vendor advisory. No public exploit code has been identified and EPSS at 0.15% (4th percentile) indicates low near-term exploitation probability, though the lack of authentication requirement warrants patching for any deployment exposing WebSocket endpoints to untrusted networks.