When curl retrieves an HTTP response, it stores the incoming headers so that they can be accessed later via the libcurl headers API. However, curl did not have a limit in how many or how large headers it would accept in a response, allowing a malicious server to stream an endless series of headers and eventually cause curl to run out of heap memory.

libcurl versions from 7.36.0 to before 7.64.0 is vulnerable to a heap buffer out-of-bounds read. The function handling incoming NTLM type-2 messages (`lib/vauth/ntlm.c:ntlm_decode_type2_target`) does not validate incoming data correctly and is subject to an integer overflow vulnerability. Using that overflow, a malicious or broken NTLM server could trick libcurl to accept a bad length + offset combination that would lead to a buffer read out-of-bounds.

1. A cookie is set using the `secure` keyword for `https://target` 2. curl is redirected to or otherwise made to speak with `http://target` (same hostname, but using clear text HTTP) using the same cookie set 3. The same cookie name is set - but with just a slash as path (`path=\"/\",`). Since this site is not secure, the cookie *should* just be ignored. 4. A bug in the path comparison logic makes curl read outside a heap buffer boundary The bug either causes a crash or it potentially makes the comparison come to the wrong conclusion and lets the clear-text site override the contents of the secure cookie, contrary to expectations and depending on the memory contents immediately following the single-byte allocation that holds the path. The presumed and correct behavior would be to plainly ignore the second set of the cookie since it was already set as secure on a secure host so overriding it on an insecure host should not be okay.

An issue in curl’s QUIC UDP receive function allows a malicious HTTP/3 server to trigger a remote denial of service against a curl or libcurl client. Because the helper function discards zero-length UDP datagrams before counting them toward the per-call packet budget, a connected QUIC peer can continuously stream empty datagrams to indefinitely stall the client.

libcurl provides the `CURLOPT_CERTINFO` option to allow applications torequest details to be returned about a server's certificate chain.Due to an erroneous function, a malicious server could make libcurl built withNSS get stuck in a never-ending busy-loop when trying to retrieve thatinformation.

Using libcurl, when a custom `Host:` header is first set for an HTTP request and a second request is subsequently done using the same *easy handle* but without the custom `Host:` header set, the second request would use stale information and pass on cookies meant for the first host in the second request. Leak them.

Due to a mistake in libcurl's WebSocket code, a malicious server can send a particularly crafted packet which makes libcurl get trapped in an endless busy-loop. There is no other way for the application to escape or exit this loop other than killing the thread/process. This might be used to DoS libcurl-using application.

When doing a second SMB request to the same host again, curl would wrongly use a data pointer pointing into already freed memory.

libcurl's ASN1 parser has this utf8asn1str() function used for parsing an ASN.1 UTF-8 string. Itcan detect an invalid field and return error. Unfortunately, when doing so it also invokes `free()` on a 4 byte localstack buffer. Most modern malloc implementations detect this error and immediately abort. Some however accept the input pointer and add that memory to its list of available chunks. This leads to the overwriting of nearby stack memory. The content of the overwrite is decided by the `free()` implementation; likely to be memory pointers and a set of flags. The most likely outcome of exploting this flaw is a crash, although it cannot be ruled out that more serious results can be had in special circumstances.

libcurl-using applications can ask for a specific client certificate to be used in a transfer. This is done with the `CURLOPT_SSLCERT` option (`--cert` with the command line tool).When libcurl is built to use the macOS native TLS library Secure Transport, an application can ask for the client certificate by name or with a file name - using the same option. If the name exists as a file, it will be used instead of by name.If the appliction runs with a current working directory that is writable by other users (like `/tmp`), a malicious user can create a file name with the same name as the app wants to use by name, and thereby trick the application to use the file based cert instead of the one referred to by name making libcurl send the wrong client certificate in the TLS connection handshake.

curl 7.21.0 to and including 7.73.0 is vulnerable to uncontrolled recursion due to a stack overflow issue in FTP wildcard match parsing.

An allocation of resources without limits or throttling vulnerability exists in curl <v7.88.0 based on the "chained" HTTP compression algorithms, meaning that a server response can be compressed multiple times and potentially with differentalgorithms. The number of acceptable "links" in this "decompression chain" wascapped, but the cap was implemented on a per-header basis allowing a maliciousserver to insert a virtually unlimited number of compression steps simply byusing many headers. The use of such a decompression chain could result in a "malloc bomb", making curl end up spending enormous amounts of allocated heap memory, or trying to and returning out of memory errors.

curl < 7.84.0 supports "chained" HTTP compression algorithms, meaning that a serverresponse can be compressed multiple times and potentially with different algorithms. The number of acceptable "links" in this "decompression chain" was unbounded, allowing a malicious server to insert a virtually unlimited number of compression steps.The use of such a decompression chain could result in a "malloc bomb", makingcurl end up spending enormous amounts of allocated heap memory, or trying toand returning out of memory errors.

A malicious server can serve excessive amounts of `Set-Cookie:` headers in a HTTP response to curl and curl < 7.84.0 stores all of them. A sufficiently large amount of (big) cookies make subsequent HTTP requests to this, or other servers to which the cookies match, create requests that become larger than the threshold that curl uses internally to avoid sending crazy large requests (1048576 bytes) and instead returns an error.This denial state might remain for as long as the same cookies are kept, match and haven't expired. Due to cookie matching rules, a server on `foo.example.com` can set cookies that also would match for `bar.example.com`, making it it possible for a "sister server" to effectively cause a denial of service for a sibling site on the same second level domain using this method.

ABB, Phoenix Contact, Schneider Electric, Siemens, WAGO - Programmable Logic Controllers, multiple versions. Researchers have found some controllers are susceptible to a denial-of-service attack due to a flood of network packets.

Suricata is a network Intrusion Detection System, Intrusion Prevention System and Network Security Monitoring engine. Prior to version 7.0.3, excessive memory use during pgsql parsing could lead to OOM-related crashes. This vulnerability is patched in 7.0.3. As workaround, users can disable the pgsql app layer parser.

The web api server on Port 8080 of ASUS HG100 firmware up to 1.05.12, which is vulnerable to Slowloris HTTP Denial of Service: an attacker can cause a Denial of Service (DoS) by sending headers very slowly to keep HTTP or HTTPS connections and associated resources alive for a long period of time. CVSS 3.0 Base score 7.4 (Availability impacts). CVSS vector: (CVSS:3.0/AV:A/AC:L/PR:N/UI:N/S:C/C:N/I:N/A:H).

When SSL Client Certificate LDAP or Certificate Revocation List Distribution Point (CRLDP) authentication profile is configured on a virtual server, undisclosed requests can cause an increase in CPU resource utilization. Note: Software versions which have reached End of Technical Support (EoTS) are not evaluated

Suricata is a network Intrusion Detection System, Intrusion Prevention System and Network Security Monitoring engine. Prior to versions 6.0.16 and 7.0.3, an attacker can craft traffic to cause Suricata to use far more CPU and memory for processing the traffic than needed, which can lead to extreme slow downs and denial of service. This vulnerability is patched in 6.0.16 or 7.0.3. Workarounds include disabling the affected protocol app-layer parser in the yaml and reducing the `stream.reassembly.depth` value helps reduce the severity of the issue.

LibHTP is a security-aware parser for the HTTP protocol. Crafted traffic can cause excessive processing time of HTTP headers, leading to denial of service. This issue is addressed in 0.5.46.

A denial-of-service vulnerability exists in the NetX IPv6 component functionality of Eclipse ThreadX NetX Duo. A specially crafted network packet of "Packet Too Big" with more than 15 different source address can lead to denial of service. An attacker can send a malicious packet to trigger this vulnerability.

Very large headers can cause resource exhaustion when parsing message. The message-parser normally reads reasonably sized chunks of the message. However, when it feeds them to message-header-parser, it starts building up "full_value" buffer out of the smaller chunks. The full_value buffer has no size limit, so large headers can cause large memory usage. It doesn't matter whether it's a single long header line, or a single header split into multiple lines. This bug exists in all Dovecot versions. Incoming mails typically have some size limits set by MTA, so even largest possible header size may still fit into Dovecot's vsz_limit. So attackers probably can't DoS a victim user this way. A user could APPEND larger mails though, allowing them to DoS themselves (although maybe cause some memory issues for the backend in general). One can implement restrictions on headers on MTA component preceding Dovecot. No publicly available exploits are known.

OpenClaw before 2026.3.31 contains an incomplete fix for CVE-2026-32062 where the voice-call component parses large WebSocket frames before start validation. Remote attackers can send oversized pre-start WebSocket frames to cause resource consumption and denial of service.

Jonathan Looney discovered that the Linux kernel default MSS is hard-coded to 48 bytes. This allows a remote peer to fragment TCP resend queues significantly more than if a larger MSS were enforced. A remote attacker could use this to cause a denial of service. This has been fixed in stable kernel releases 4.4.182, 4.9.182, 4.14.127, 4.19.52, 5.1.11, and is fixed in commits 967c05aee439e6e5d7d805e195b3a20ef5c433d6 and 5f3e2bf008c2221478101ee72f5cb4654b9fc363.

IBM WebSphere Application Server Liberty 17.0.0.3 through 24.0.0.4 is vulnerable to a denial of service, caused by sending a specially crafted request. A remote attacker could exploit this vulnerability to cause the server to consume memory resources. IBM X-Force ID: 280400.

It was found that the fix for CVE-2018-14648 in 389-ds-base, versions 1.4.0.x before 1.4.0.17, was incorrectly applied in RHEL 7.5. An attacker would still be able to provoke excessive CPU consumption leading to a denial of service.

Jetty is a Java based web server and servlet engine. An HTTP/2 SSL connection that is established and TCP congested will be leaked when it times out. An attacker can cause many connections to end up in this state, and the server may run out of file descriptors, eventually causing the server to stop accepting new connections from valid clients. The vulnerability is patched in 9.4.54, 10.0.20, 11.0.20, and 12.0.6.

Apache Traffic Server is vulnerable to HTTP/2 setting flood attacks. Earlier versions of Apache Traffic Server didn't limit the number of setting frames sent from the client using the HTTP/2 protocol. Users should upgrade to Apache Traffic Server 7.1.7, 8.0.4, or later versions.

Allocation of Resources Without Limits or Throttling vulnerability in Apache Tomcat. This issue affects Apache Tomcat: from 11.0.0-M1 through 11.0.21, from 10.1.0-M1 through 10.1.54, from 9.0.0.M1 through 9.0.117. Older, unsupported versions may also be affected. Users are recommended to upgrade to version [FIXED_VERSION], which fixes the issue.

Versions of the package @eslint/plugin-kit before 0.2.3 are vulnerable to Regular Expression Denial of Service (ReDoS) due to improper input sanitization. An attacker can increase the CPU usage and crash the program by exploiting this vulnerability.

Amazon Ion is a Java implementation of the Ion data notation. Prior to version 1.10.5, a potential denial-of-service issue exists in `ion-java` for applications that use `ion-java` to deserialize Ion text encoded data, or deserialize Ion text or binary encoded data into the `IonValue` model and then invoke certain `IonValue` methods on that in-memory representation. An actor could craft Ion data that, when loaded by the affected application and/or processed using the `IonValue` model, results in a `StackOverflowError` originating from the `ion-java` library. The patch is included in `ion-java` 1.10.5. As a workaround, do not load data which originated from an untrusted source or that could have been tampered with.

For unspecified traffic patterns, BIG-IP AFM IPS engine may spend an excessive amount of time matching the traffic against signatures, resulting in Traffic Management Microkernel (TMM) restarting and traffic disruption.  Note: Software versions which have reached End of Technical Support (EoTS) are not evaluated

An Allocation of Resources Without Limits or Throttling vulnerability in the kernel of Juniper Networks Junos OS Evolved allows an unauthenticated, network-based attacker to cause a Denial of Service (DoS). If a high rate of specific valid packets are processed by the routing engine (RE) this will lead to a loss of connectivity of the RE with other components of the chassis and thereby a complete and persistent system outage. Please note that a carefully designed lo0 firewall filter will block or limit these packets which should prevent this issue from occurring. The following log messages can be seen when this issue occurs: <host> kernel: nf_conntrack: nf_conntrack: table full, dropping packet This issue affects Juniper Networks Junos OS Evolved: * All versions earlier than 20.4R3-S7-EVO; * 21.2R1-EVO and later versions; * 21.4-EVO versions earlier than 21.4R3-S5-EVO; * 22.1-EVO versions earlier than 22.1R3-S2-EVO; * 22.2-EVO versions earlier than 22.2R3-EVO; * 22.3-EVO versions earlier than 22.3R2-EVO; * 22.4-EVO versions earlier than 22.4R2-EVO.

In Micrometer, it is possible for a user to provide specially crafted HTTP requests that may cause a denial-of-service (DoS) condition. Affected versions: micrometer-core 1.16.0 through 1.16.5; 1.15.0 through 1.15.11; 1.14.0 through 1.14.15; 1.13.0 through 1.13.18; 1.9.0 through 1.9.17. micrometer-jetty11 1.16.0 through 1.16.5; 1.15.0 through 1.15.11; 1.14.0 through 1.14.15; 1.13.0 through 1.13.18. micrometer-jetty12 1.16.0 through 1.16.5; 1.15.0 through 1.15.11; 1.14.0 through 1.14.15; 1.13.0 through 1.13.18.

Specific IPv6 DHCP packets received by the jdhcpd daemon will cause a memory resource consumption issue to occur on a Junos OS device using the jdhcpd daemon configured to respond to IPv6 requests. Once started, memory consumption will eventually impact any IPv4 or IPv6 request serviced by the jdhcpd daemon, thus creating a Denial of Service (DoS) condition to clients requesting and not receiving IP addresses. Additionally, some clients which were previously holding IPv6 addresses will not have their IPv6 Identity Association (IA) address and network tables agreed upon by the jdhcpd daemon after the failover event occurs, which leads to more than one interface, and multiple IP addresses, being denied on the client. Affected releases are Juniper Networks Junos OS: 17.4 versions prior to 17.4R2; 18.1 versions prior to 18.1R2.

Cloudflare Quiche (through version 0.19.1/0.20.0) was affected by an unlimited resource allocation vulnerability causing rapid increase of memory usage of the system running quiche server or client. A remote attacker could take advantage of this vulnerability by repeatedly sending an unlimited number of 1-RTT CRYPTO frames after previously completing the QUIC handshake. Exploitation was possible for the duration of the connection which could be extended by the attacker.  quiche 0.19.2 and 0.20.1 are the earliest versions containing the fix for this issue.

An issue was discovered in MediaWiki through 1.38.1. The lemma length of a Wikibase lexeme is currently capped at a thousand characters. Unfortunately, this length is not validated, allowing much larger lexemes to be created, which introduces various denial-of-service attack vectors within the Wikibase and WikibaseLexeme extensions. This is related to Special:NewLexeme and Special:NewProperty.

CryptPad 2025.3.1 allows unbounded WebSocket frame flood. A remote, unauthenticated attacker can significantly degrade or deny service for all users of a CryptPad instance. Fixed in 2026.2.2.

Resolver caches and authoritative zone databases that hold significant numbers of RRs for the same hostname (of any RTYPE) can suffer from degraded performance as content is being added or updated, and also when handling client queries for this name. This issue affects BIND 9 versions 9.11.0 through 9.11.37, 9.16.0 through 9.16.50, 9.18.0 through 9.18.27, 9.19.0 through 9.19.24, 9.11.4-S1 through 9.11.37-S1, 9.16.8-S1 through 9.16.50-S1, and 9.18.11-S1 through 9.18.27-S1.

An Environment (CWE-2) vulnerability exists in SoMachine Basic, all versions, and Modicon M221(all references, all versions prior to firmware V1.10.0.0) which could cause cycle time impact when flooding the M221 ethernet interface while the Ethernet/IP adapter is activated.

The ppp decapsulator in tcpdump 4.9.3 can be convinced to allocate a large amount of memory.

CNCF Envoy through 1.13.0 may consume excessive amounts of memory when proxying HTTP/1.1 requests or responses with many small (i.e. 1 byte) chunks.

basic-ftp is an FTP client for Node.js. Versions prior to 5.3.0 are vulnerable to denial of service through unbounded memory growth while processing directory listings from a remote FTP server. A malicious or compromised server can send an extremely large or never-ending listing response to `Client.list()`, causing the client process to consume memory until it becomes unstable or crashes. Version 5.3.0 fixes the issue.

HashiCorp Consul and Consul Enterprise up to 1.6.2 HTTP/RPC services allowed unbounded resource usage, and were susceptible to unauthenticated denial of service. Fixed in 1.6.3.

By design, BIND is intended to limit the number of TCP clients that can be connected at any given time. The number of allowed connections is a tunable parameter which, if unset, defaults to a conservative value for most servers. Unfortunately, the code which was intended to limit the number of simultaneous connections contained an error which could be exploited to grow the number of simultaneous connections beyond this limit. Versions affected: BIND 9.9.0 -> 9.10.8-P1, 9.11.0 -> 9.11.6, 9.12.0 -> 9.12.4, 9.14.0. BIND 9 Supported Preview Edition versions 9.9.3-S1 -> 9.11.5-S3, and 9.11.5-S5. Versions 9.13.0 -> 9.13.7 of the 9.13 development branch are also affected. Versions prior to BIND 9.9.0 have not been evaluated for vulnerability to CVE-2018-5743.

Clients using DNS-over-HTTPS (DoH) can exhaust a DNS resolver's CPU and/or memory by flooding it with crafted valid or invalid HTTP/2 traffic. This issue affects BIND 9 versions 9.18.0 through 9.18.32, 9.20.0 through 9.20.4, 9.21.0 through 9.21.3, and 9.18.11-S1 through 9.18.32-S1.

In version 0.3.32 of open-webui/open-webui, the absence of authentication mechanisms allows any unauthenticated attacker to access the `api/v1/utils/code/format` endpoint. If a malicious actor sends a POST request with an excessively high volume of content, the server could become completely unresponsive. This could lead to severe performance issues, causing the server to become unresponsive or experience significant degradation, ultimately resulting in service interruptions for legitimate users.

Fastify node module before 0.38.0 is vulnerable to a denial-of-service attack by sending a request with "Content-Type: application/json" and a very large payload.

@fastify/accepts-serializer cached serializer-selection results keyed by the request Accept header without a size limit or eviction policy. A remote unauthenticated client could send many distinct but matching Accept header variants to make the cache grow unbounded, eventually exhausting the Node.js heap and crashing the process. Versions <= 6.0.3 are affected. Update to 6.0.4 or later, which bounds the cache via an LRU with a default size of 100 entries, configurable through the new cacheSize plugin option.