vLLM is an inference and serving engine for large language models. Prior to 0.24.0, the input_audio handling path for /v1/chat/completions calls AudioMediaIO.load_bytes or AudioMediaIO.load_file without passing VLLM_MAX_AUDIO_DECODE_DURATION_S to the shared audio decoder. An unauthenticated client can therefore submit a small compressed audio input that expands into a very large float32 PCM allocation, bypassing the duration guard already used by /v1/audio/transcriptions and causing an out-of-memory worker crash. Inline data URLs reach this path without being bounded by VLLM_AUDIO_FETCH_TIMEOUT. The issue affects deployments serving an audio-capable model, and authentication changes only the deployment-specific reachability. This issue is fixed in version 0.24.0.
RabbitMQ amqp091-go is a Go AMQP 0.9.1 client. Prior to 1.13.0, Channel.dispatch in channel.go, confirms.confirm in confirms.go, and Connection.dispatch0 in connection.go synchronously send publisher confirmations, flow-control events, consumer cancellations, returned messages, including NotifyConfirm events and connection block notifications, to application-provided channels. If a listener channel is unbuffered, full, or not drained promptly, the sole reader goroutine blocks and stops processing frames, acknowledgments, deliveries, and heartbeats. Broker-driven event bursts can therefore cause connection stalls, missed heartbeats, deadlocks, and disconnection. This issue is fixed in version 1.13.0.
RabbitMQ amqp091-go is a Go AMQP 0.9.1 client. Prior to 1.13.0, Connection.openTune in connection.go accepts a server-advertised FrameMax below the AMQP frameMinSize value of 4096 bytes because the connection negotiation loop does not enforce the protocol minimum. A malicious or compromised AMQP broker can therefore advertise an extremely small FrameMax, causing later client publications to be fragmented into excessive numbers of frames and write operations. This can consume CPU and stall the client or its host. This issue is fixed in version 1.13.0.
CVE-2026-85501 affects NLnet Labs Unbound DNS resolver up to version 1.26.0 and involves allocation of resources without limits or throttling, leading to potential service degradation. The vulnerability arises from algorithmic complexity attacks on DNSSEC validation, including multiple attack vectors such as TagTrap, DelegationTrap, NsecTrap, and AdditionalTrap. These exploit excessive or malformed DNSSEC-related records to cause high computational overhead and resource exhaustion. The CVSS score is 5.3, indicating a medium severity impact primarily on availability. No known exploits are reported in the wild. No official patch or vendor advisory is provided in the input data, so patch status is not confirmed.
CVE-2026-80225 is a medium severity denial-of-service vulnerability in NLnet Labs Unbound up to version 1.26.0. It arises from allocation of resources without limits or throttling during TCP/DoT reading, allowing an attacker to monopolize a worker's event loop by streaming many distinct uncached DNS names. This causes service degradation but does not affect confidentiality or integrity.
Skipper is an HTTP router and reverse proxy for service composition. Prior to 0.26.22, Handler in dataclients/kubernetes/admission/admission.go passes the body of requests to the Kubernetes admission endpoint at :9443/admission directly to io.ReadAll(r.Body) without a size limit. An attacker with in-cluster network access and a valid Kubernetes client certificate can send a very large body that causes unbounded memory allocation and an out-of-memory termination of the Skipper process. The disruption is limited to Ingress and RouteGroup admission rather than pod creation or unrelated admission controllers, and Kubernetes normally restarts the process. This issue is fixed in version 0.26.22.
docx4j is an open source Java library for creating, editing, and saving OpenXML packages, including DOCX, PPTX, and XLSX files. Prior to 11.5.14, PropertyResolver and adjacent helpers recursively follow the WordprocessingML w:basedOn style inheritance chain without cycle detection. A well-formed DOCX containing mutually based styles causes unbounded recursion in PropertyResolver.fillPPrStack and related effective-style resolution paths, resulting in StackOverflowError. Server-side conversion and table-of-contents processing of an untrusted document can terminate a worker thread, degrade a thread pool, or deny service, although isolation in disposable workers or safe containment of StackOverflowError can reduce the practical effect. The fix adds cyclic-style tracking and CyclicStylesException handling. This issue is fixed in version 11.5.14.
CVE-2026-50276 is a vulnerability in DataDog's dd-trace-rb Ruby client library prior to version 2.32.0. The issue involves the W3C baggage extraction process, which does not enforce configured limits on the number or size of baggage items during extraction, although these limits apply during injection. This allows a remote unauthenticated attacker to send a specially crafted baggage HTTP header with many key-value pairs or a very large value, causing unbounded CPU and memory consumption. This can lead to denial of service in HTTP services using the affected tracers with baggage propagation enabled. The vulnerability is fixed in dd-trace-rb version 2.32.0.
CVE-2026-50270 is a vulnerability in DataDog's dd-trace-java APM client for Java versions prior to 1.62.0. The issue arises because the W3C baggage extraction process does not enforce configured limits on the number or size of baggage items during extraction, although these limits are enforced during injection. This allows a remote unauthenticated attacker to send a specially crafted HTTP baggage header with many key-value pairs or a very large value, causing unbounded CPU and memory consumption. This can lead to denial of service in HTTP services using the default baggage propagation style. The vulnerability is fixed in version 1.62.0.
webtransport-go is an implementation of the WebTransport protocol. Prior to 0.11.1, Session.parseNextCapsule() in session.go skips an unknown WebTransport capsule on the HTTP/3 request stream by calling io.ReadAll on the capsule reader, retaining the complete declared capsule body in memory. A malicious peer can send a large unknown capsule and cause a client or server to allocate memory for the full payload; QUIC flow control does not bound the total retained allocation because reading advances the flow-control window while the received bytes remain in memory. The resulting memory and resource exhaustion can disrupt or crash the affected process. This issue is fixed in version 0.11.1.
Description
The DRPC server kept a map from function name to request queue and created an entry the first time a
function name was seen. No code path ever removed an entry: request cleanup removed the request from its
queue, and the shutdown path drained queues, but the queue object and its map entry remained for the life of
the process.
Function names come from the client and are not constrained to functions any topology has registered, so the
number of retained entries is bounded only by the number of distinct names an attacker chooses to send, and
each retained entry holds the name itself. `drpc.authorizer` is unset by default, so no credentials are
required to reach the endpoint.
The retained state is permanent rather than a transient load spike, so the effect accumulates until the DRPC
server exhausts its heap.
Mitigation
Upgrade to 3.1.0, where a function's queue is removed once nothing is waiting in it.
Users who cannot upgrade immediately should configure `drpc.authorizer` so that only trusted principals can
reach the DRPC endpoints, and should ensure the DRPC ports are not reachable from untrusted networks.
Credit
The ASF -- found using Claude agents to study the security of open-source projects, validated and reported by Apache Storm.
Hoverfly is an open source API simulation tool. Prior to version 1.12.8, remote post-serve actions use `http.DefaultClient` without any timeout configuration. When the remote endpoint is unreachable or intentionally slow (accepts TCP connection but never responds), each triggered proxy request spawns a goroutine that blocks indefinitely on `http.DefaultClient.Do()`. An attacker can cause unbounded goroutine accumulation leading to memory exhaustion and process crash (OOM kill). Unlike local post-serve action execution, this requires no binary execution, only a URL pointing to a non-responsive endpoint. Version 1.12.8 patches the issue.
OpenTelemetry eBPF Profiler is a production-scale agent for profiling applications across multiple programming languages. Starting in version 0.0.202527 and prior to version 0.0.202622, an unprivileged process can cause the profiler to open a nonregular mapping file, such as a FIFO, and block indefinitely, preventing further ELF analysis and causing a denial of service. Version 0.0.202622 contains a patch. No known workarounds are available.
AirSane is a SANE frontend, and a scanner server that supports Apple's AirScan protocol. Versions prior to 0.4.12 have a vulnerability in the custom HTTP server implementation of AirSane that allows a remote unauthenticated attacker to cause a Denial of Service (DoS) via memory exhaustion (OOM). In httpserver.cpp, the HttpServer::Request::content function reads the Content-Length header and directly passes this value to std::string::resize() without any upper-bound validation or safe parsing. An attacker can send an HTTP POST request with an artificially large Content-Length value. This forces the daemon to attempt allocating gigabytes of memory, resulting in a std::bad_alloc exception and immediately crashing the AirSane process. Additionally, providing non-numeric characters in the Content-Length header leads to undefined behavior (NaN to integer conversion) due to the lack of error handling during header parsing. Version 0.4.12 patches the issue.
CVE-2026-45769 is a high-severity vulnerability in the Suricata network security engine affecting versions prior to 7.0.16 and between 8.0.0 and before 8.0.5. The issue involves uncontrolled resource consumption due to the IKEv2 parser state growing without bounds when storing client transforms. This can lead to excessive memory use and potential denial of service. The vulnerability is fixed in Suricata versions 7.0.16 and 8.0.5. Mitigations include disabling IKE application-layer parsing if not needed or using a rule to bypass IKE flows after a certain number of packets.
CVE-2026-45768 is a high-severity vulnerability in Suricata, a network intrusion detection and prevention system. Versions from 8.0.0 up to but not including 8.0.5 are affected. The issue involves uncontrolled resource consumption due to LDAP transaction state storing an unbounded number of responses, which can lead to excessive memory use and potential denial of service. The vulnerability arises because LDAP can be processed over UDP, allowing crafted traffic to trigger this condition. Version 8.0.5 includes a fix for this problem. As a mitigation, disabling LDAP application-layer parsing when not needed or using a detection rule to alert on excessive LDAP responses is recommended.
CVE-2026-45766 is a high-severity vulnerability in Suricata, a network intrusion detection and prevention system. Versions prior to 7.0.16 and 8.0.5 have insufficient bounds on certain NFS parser state structures. This flaw allows crafted NFS traffic to cause excessive memory consumption, potentially leading to denial of service. Fixed in versions 7.0.16 and 8.0.5. Disabling NFS application-layer parsing is a recommended workaround if that functionality is not required.
CVE-2026-45765 is a high-severity vulnerability in Suricata, a network intrusion detection and prevention system. Prior to versions 7.0.16 and 8.0.5, Suricata's DNP3 protocol reassembly could consume excessive memory due to insufficient parser-level bounds, potentially leading to denial of service. The issue is fixed in versions 7.0.16 and 8.0.5. As a mitigation, users can disable DNP3 if not needed or limit the stream reassembly depth.
CVE-2026-45763 is a vulnerability in Suricata versions 8.0.0 through 8.0.4 where the Lua sandbox memory limit was not consistently enforced. This flaw allows certain Lua allocation patterns to exceed the configured memory limit, potentially causing resource exhaustion. The issue requires Lua rule execution to be enabled and an affected Lua script or rule to be loaded. The vulnerability is fixed in version 8.0.5. A workaround is to disable Lua rules unless they are required.