CWE-409

.NET and Visual Studio Denial of Service Vulnerability

Package jose aims to provide an implementation of the Javascript Object Signing and Encryption set of standards. An attacker could send a JWE containing compressed data that used large amounts of memory and CPU when decompressed by Decrypt or DecryptMulti. Those functions now return an error if the decompressed data would exceed 250kB or 10x the compressed size (whichever is larger). This vulnerability has been patched in versions 4.0.1, 3.0.3 and 2.6.3.

The HTTP client module superagent is vulnerable to ZIP bomb attacks. In a ZIP bomb attack, the HTTP server replies with a compressed response that becomes several magnitudes larger once uncompressed. If a client does not take special care when processing such responses, it may result in excessive CPU and/or memory consumption. An attacker might exploit such a weakness for a DoS attack. To exploit this the attacker must control the location (URL) that superagent makes a request to.

Envoy is a cloud-native high-performance proxy. In versions prior to 1.22.1 secompressors accumulate decompressed data into an intermediate buffer before overwriting the body in the decode/encodeBody. This may allow an attacker to zip bomb the decompressor by sending a small highly compressed payload. Maliciously constructed zip files may exhaust system memory and cause a denial of service. Users are advised to upgrade. Users unable to upgrade may consider disabling decompression.

Pillow before 9.2.0 performs Improper Handling of Highly Compressed GIF Data (Data Amplification).

gosaml2 is a Pure Go implementation of SAML 2.0. SAML Service Providers using this library for SAML authentication support are likely susceptible to Denial of Service attacks. A bug in this library enables attackers to craft a `deflate`-compressed request which will consume significantly more memory during processing than the size of the original request. This may eventually lead to memory exhaustion and the process being killed. The maximum compression ratio achievable with `deflate` is 1032:1, so by limiting the size of bodies passed to gosaml2, limiting the rate and concurrency of calls, and ensuring that lots of memory is available to the process it _may_ be possible to help Go's garbage collector "keep up". Implementors are encouraged not to rely on this. This issue is fixed in version 0.9.0.

The scrapy/scrapy project is vulnerable to XML External Entity (XXE) attacks due to the use of lxml.etree.fromstring for parsing untrusted XML data without proper validation. This vulnerability allows attackers to perform denial of service attacks, access local files, generate network connections, or circumvent firewalls by submitting specially crafted XML data.

HashiCorp Nomad and Nomad Enterprise 1.2.15 up to 1.3.8, and 1.4.3 jobs using a maliciously compressed artifact stanza source can cause excessive disk usage. Fixed in 1.2.16, 1.3.9, and 1.4.4.

The Apollo Router is a graph router written in Rust to run a federated supergraph that uses Apollo Federation. Versions 0.9.5 until 1.40.2 are subject to a Denial-of-Service (DoS) type vulnerability. When receiving compressed HTTP payloads, affected versions of the Router evaluate the `limits.http_max_request_bytes` configuration option after the entirety of the compressed payload is decompressed. If affected versions of the Router receive highly compressed payloads, this could result in significant memory consumption while the compressed payload is expanded. Router version 1.40.2 has a fix for the vulnerability. Those who are unable to upgrade may be able to implement mitigations at proxies or load balancers positioned in front of their Router fleet (e.g. Nginx, HAProxy, or cloud-native WAF services) by creating limits on HTTP body upload size.

In h2oai/h2o-3 version 3.46.0.2, a vulnerability exists where uploading and repeatedly parsing a large GZIP file can cause a denial of service. The server becomes unresponsive due to memory exhaustion and a large number of concurrent slow-running jobs. This issue arises from the improper handling of highly compressed data, leading to significant data amplification.

Improper Handling of Highly Compressed Data (Compression Bomb) vulnerability in Erlang OTP ssh (ssh_transport modules) allows Denial of Service via Resource Depletion. The SSH transport layer advertises legacy zlib compression by default and inflates attacker-controlled payloads pre-authentication without any size limit, enabling reliable memory exhaustion DoS. Two compression algorithms are affected: * zlib: Activates immediately after key exchange, enabling unauthenticated attacks * zlib@openssh.com: Activates post-authentication, enabling authenticated attacks Each SSH packet can decompress ~255 MB from 256 KB of wire data (1029:1 amplification ratio). Multiple packets can rapidly exhaust available memory, causing OOM kills in memory-constrained environments. This vulnerability is associated with program files lib/ssh/src/ssh_transport.erl and program routines ssh_transport:decompress/2, ssh_transport:handle_packet_part/4. This issue affects OTP from OTP 17.0 until OTP 28.4.1, 27.3.4.9 and 26.2.5.18 corresponding to ssh from 3.0.1 until 5.5.1, 5.2.11.6 and 5.1.4.14.

The undici WebSocket client is vulnerable to a denial-of-service attack via unbounded memory consumption during permessage-deflate decompression. When a WebSocket connection negotiates the permessage-deflate extension, the client decompresses incoming compressed frames without enforcing any limit on the decompressed data size. A malicious WebSocket server can send a small compressed frame (a "decompression bomb") that expands to an extremely large size in memory, causing the Node.js process to exhaust available memory and crash or become unresponsive. The vulnerability exists in the PerMessageDeflate.decompress() method, which accumulates all decompressed chunks in memory and concatenates them into a single Buffer without checking whether the total size exceeds a safe threshold.

An Out-Of-Memory (OOM) vulnerability exists in the `ollama` server version 0.3.14. This vulnerability can be triggered when a malicious API server responds with a gzip bomb HTTP response, leading to the `ollama` server crashing. The vulnerability is present in the `makeRequestWithRetry` and `getAuthorizationToken` functions, which use `io.ReadAll` to read the response body. This can result in excessive memory usage and a Denial of Service (DoS) condition.

A vulnerability in the binary-husky/gpt_academic repository, as of commit git 3890467, allows an attacker to crash the server by uploading a specially crafted zip bomb. The server decompresses the uploaded file and attempts to load it into memory, which can lead to an out-of-memory crash. This issue arises due to improper input validation when handling compressed file uploads.

Net::IMAP implements Internet Message Access Protocol (IMAP) client functionality in Ruby. Starting in version 0.3.2 and prior to versions 0.3.8, 0.4.19, and 0.5.6, there is a possibility for denial of service by memory exhaustion in `net-imap`'s response parser. At any time while the client is connected, a malicious server can send can send highly compressed `uid-set` data which is automatically read by the client's receiver thread. The response parser uses `Range#to_a` to convert the `uid-set` data into arrays of integers, with no limitation on the expanded size of the ranges. Versions 0.3.8, 0.4.19, 0.5.6, and higher fix this issue. Additional details for proper configuration of fixed versions and backward compatibility are available in the GitHub Security Advisory.

Netty is an asynchronous event-driven network application framework for rapid development of maintainable high performance protocol servers & clients. In netty-codec-compression versions 4.1.124.Final and below, and netty-codec versions 4.2.4.Final and below, when supplied with specially crafted input, BrotliDecoder and certain other decompression decoders will allocate a large number of reachable byte buffers, which can lead to denial of service. BrotliDecoder.decompress has no limit in how often it calls pull, decompressing data 64K bytes at a time. The buffers are saved in the output list, and remain reachable until OOM is hit. This is fixed in versions 4.1.125.Final of netty-codec and 4.2.5.Final of netty-codec-compression.

urllib3 is a user-friendly HTTP client library for Python. Starting in version 1.0 and prior to 2.6.0, the Streaming API improperly handles highly compressed data. urllib3's streaming API is designed for the efficient handling of large HTTP responses by reading the content in chunks, rather than loading the entire response body into memory at once. When streaming a compressed response, urllib3 can perform decoding or decompression based on the HTTP Content-Encoding header (e.g., gzip, deflate, br, or zstd). The library must read compressed data from the network and decompress it until the requested chunk size is met. Any resulting decompressed data that exceeds the requested amount is held in an internal buffer for the next read operation. The decompression logic could cause urllib3 to fully decode a small amount of highly compressed data in a single operation. This can result in excessive resource consumption (high CPU usage and massive memory allocation for the decompressed data.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in Apache Seata (incubating). This issue affects Apache Seata (incubating): through <=2.2.0. Users are recommended to upgrade to version 2.3.0, which fixes the issue.

urllib3 is an HTTP client library for Python. urllib3's streaming API is designed for the efficient handling of large HTTP responses by reading the content in chunks, rather than loading the entire response body into memory at once. urllib3 can perform decoding or decompression based on the HTTP `Content-Encoding` header (e.g., `gzip`, `deflate`, `br`, or `zstd`). When using the streaming API, the library decompresses only the necessary bytes, enabling partial content consumption. Starting in version 1.22 and prior to version 2.6.3, for HTTP redirect responses, the library would read the entire response body to drain the connection and decompress the content unnecessarily. This decompression occurred even before any read methods were called, and configured read limits did not restrict the amount of decompressed data. As a result, there was no safeguard against decompression bombs. A malicious server could exploit this to trigger excessive resource consumption on the client. Applications and libraries are affected when they stream content from untrusted sources by setting `preload_content=False` when they do not disable redirects. Users should upgrade to at least urllib3 v2.6.3, in which the library does not decode content of redirect responses when `preload_content=False`. If upgrading is not immediately possible, disable redirects by setting `redirect=False` for requests to untrusted source.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion. cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2. This issue affects cowlib from 0.1.0 before 2.16.1.

Unfurl before 2026.04 contains an unbounded zlib decompression vulnerability in parse_compressed.py that allows remote attackers to cause denial of service. Attackers can submit highly compressed payloads via URL parameters to the /json/visjs endpoint that expand to gigabytes, exhausting server memory and crashing the service.

A flaw was found in Keycloak. An unauthenticated remote attacker can trigger an application level Denial of Service (DoS) by sending a highly compressed SAMLRequest through the SAML Redirect Binding. The server fails to enforce size limits during DEFLATE decompression, leading to an OutOfMemoryError (OOM) and subsequent process termination. This vulnerability allows an attacker to disrupt the availability of the service.

NATS-Server is a High-Performance server for NATS.io, a cloud and edge native messaging system. The WebSockets handling of NATS messages handles compressed messages via the WebSockets negotiated compression. Prior to versions 2.11.2 and 2.12.3, the implementation bound the memory size of a NATS message but did not independently bound the memory consumption of the memory stream when constructing a NATS message which might then fail validation for size reasons. An attacker can use a compression bomb to cause excessive memory consumption, often resulting in the operating system terminating the server process. The use of compression is negotiated before authentication, so this does not require valid NATS credentials to exploit. The fix, present in versions 2.11.2 and 2.12.3, was to bounds the decompression to fail once the message was too large, instead of continuing on. The vulnerability only affects deployments which use WebSockets and which expose the network port to untrusted end-points.

A denial of service (DoS) condition was discovered in GitLab CE/EE affecting all versions from 13.2.4 before 16.10.6, 16.11 before 16.11.3, and 17.0 before 17.0.1. By leveraging this vulnerability an attacker could create a DoS condition by sending crafted API calls.

Turms AI-Serving module v0.10.0-SNAPSHOT and earlier contains an image decompression bomb denial of service vulnerability. The ExtendedOpenCVImage class in ai/djl/opencv/ExtendedOpenCVImage.java loads images using OpenCV's imread() function without validating dimensions or pixel count before decompression. An attacker can upload a specially crafted compressed image file (e.g., PNG) that is small when compressed but expands to gigabytes of memory when loaded. This causes immediate memory exhaustion, OutOfMemoryError, and service crash. No authentication is required if the OCR service is publicly accessible. Multiple requests can completely deny service availability.

Chall-Manager is a platform-agnostic system able to start Challenges on Demand of a player. When decoding a scenario (i.e. a zip archive), the size of the decoded content is not checked, potentially leading to zip bombs decompression. Exploitation does not require authentication nor authorization, so anyone can exploit it. It should nonetheless not be exploitable as it is highly recommended to bury Chall-Manager deep within the infrastructure due to its large capabilities, so no users could reach the system. Patch has been implemented by commit 14042aa and shipped in v0.1.4.

This vulnerability allows any authenticated user to cause the server to consume very large amounts of disk space when extracting a Zip Bomb. If user import is enabled (which is the default setting), any registered user can upload an archive for importing. The code uses the yauzl library for reading the archive. The yauzl library does not contain any mechanism to detect or prevent extraction of a Zip Bomb https://en.wikipedia.org/wiki/Zip_bomb . Therefore, when using the User Import functionality with a Zip Bomb, PeerTube will try extracting the archive which will cause a disk space resource exhaustion.

kin-openapi is a Go project for handling OpenAPI files. Prior to 0.131.0, when validating a request with a multipart/form-data schema, if the OpenAPI schema allows it, an attacker can upload a crafted ZIP file (e.g., a ZIP bomb), causing the server to consume all available system memory. The root cause comes from the ZipFileBodyDecoder, which is registered automatically by the module (contrary to what the documentation says). This vulnerability is fixed in 0.131.0.

HashiCorp go-getter up to 1.6.2 and 2.1.1 is vulnerable to decompression bombs. Fixed in 1.7.0 and 2.2.0.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in wojtekmach Req allows attacker-controlled HTTP servers to exhaust memory in a Req client via decompression-bomb response bodies. Req's default response pipeline includes Req.Steps.decode_body/1 and Req.Steps.decompress_body/1 in lib/req/steps.ex. decode_body/1 dispatches on the server-supplied content-type (or URL extension) and calls :zip.extract(body, [:memory]) for application/zip, :erl_tar.extract({:binary, body}, [:memory]) for application/x-tar, and :erl_tar.extract({:binary, body}, [:memory, :compressed]) for application/gzip / .tgz. Each returns the full decompressed archive contents as a [{name, bytes}] list in memory, with no per-entry or total size cap. decompress_body/1 walks the content-encoding header and chains :zlib/:brotli/:ezstd decoders, so a response advertising content-encoding: gzip, gzip, gzip inflates through multiple layers without bound. Both steps are enabled by default, no caller opt-in is required, and the attacker controls the content-type and content-encoding headers on their own server (or on any host reached via Req's automatic redirect following). A sub-megabyte response can expand to multiple gigabytes on the victim, crashing the BEAM process. This issue affects req: from 0.1.0 before 0.6.1.

Protocol::HTTP2 versions before 1.13 for Perl is vulnerable to a HTTP/2 Bomb. Protocol::HTTP2's inbound HPACK path has no header-list size limit, so a small HTTP/2 request can expand into large server memory (the "HTTP/2 bomb"). The headers_decode method materialises a full key+value copy per indexed reference with no running size check, and the stream_header_block_add method appends (since version 1.12) every CONTINUATION frame to the per-stream buffer unbounded. MAX_HEADER_LIST_SIZE (default 65536) is advertised in SETTINGS but never consulted on decode. It is absent from the decoder and from the :limits export tag.

GuardDog is a CLI tool to identify malicious PyPI packages. Prior to 2.7.1, GuardDog's safe_extract() function does not validate decompressed file sizes when extracting ZIP archives (wheels, eggs), allowing attackers to cause denial of service through zip bombs. A malicious package can consume gigabytes of disk space from a few megabytes of compressed data. This vulnerability is fixed in 2.7.1.

cpp-httplib is a C++11 single-file header-only cross platform HTTP/HTTPS library. Prior to 0.35.0, cpp-httplib (httplib.h) does not enforce Server::set_payload_max_length() on the decompressed request body when using HandlerWithContentReader (streaming ContentReader) with Content-Encoding: gzip (or other supported encodings). A small compressed payload can expand beyond the configured payload limit and be processed by the application, enabling a payload size limit bypass and potential denial of service (CPU/memory exhaustion). This vulnerability is fixed in 0.35.0.

psd-tools is a Python package for working with Adobe Photoshop PSD files. Prior to version 1.12.2, when a PSD file contains malformed RLE-compressed image data (e.g. a literal run that extends past the expected row size), decode_rle() raises ValueError which propagated all the way to the user, crashing psd.composite() and psd-tools export. decompress() already had a fallback that replaces failed channels with black pixels when result is None, but it never triggered because the ValueError from decode_rle() was not caught. The fix in version 1.12.2 wraps the decode_rle() call in a try/except so the existing fallback handles the error gracefully.

MobSF is a mobile application security testing tool used. Typically, MobSF is deployed on centralized internal or cloud-based servers that also host other security tools and web applications. Access to the MobSF web interface is often granted to internal security teams, audit teams, and external vendors. MobSF provides a feature that allows users to upload ZIP files for static analysis. Upon upload, these ZIP files are automatically extracted and stored within the MobSF directory. However, in versions up to and including 4.3.2, this functionality lacks a check on the total uncompressed size of the ZIP file, making it vulnerable to a ZIP of Death (zip bomb) attack. Due to the absence of safeguards against oversized extractions, an attacker can craft a specially prepared ZIP file that is small in compressed form but expands to a massive size upon extraction. Exploiting this, an attacker can exhaust the server's disk space, leading to a complete denial of service (DoS) not just for MobSF, but also for any other applications or websites hosted on the same server. This vulnerability can lead to complete server disruption in an organization which can affect other internal portals and tools too (which are hosted on the same server). If some organization has created their customized cloud based mobile security tool using MobSF core then an attacker can exploit this vulnerability to crash their servers. Commit 6987a946485a795f4fd38cebdb4860b368a1995d fixes this issue. As an additional mitigation, it is recommended to implement a safeguard that checks the total uncompressed size of any uploaded ZIP file before extraction. If the estimated uncompressed size exceeds a safe threshold (e.g., 100 MB), MobSF should reject the file and notify the user.

Mattermost versions 10.1.x <= 10.1.2, 10.0.x <= 10.0.2, 9.11.x <= 9.11.4, 9.5.x <= 9.5.12 fail to limit the file size for slack import file uploads which allows a user to cause a DoS via zip bomb by importing data in a team they are a team admin.

Versions of the package exifreader before 4.39.0 are vulnerable to Improper Handling of Highly Compressed Data (Data Amplification) due to decompressing PNG zTXt metadata without enforcing a built-in maximum decompressed output size. When asynchronous parsing is enabled, a crafted PNG file containing a highly compressed zTXt chunk can cause ExifReader to materialize a disproportionately large Comment value in memory.

urllib3 is an HTTP client library for Python. From 2.6.0 to before 2.7.0, urllib3 could decompress the whole response instead of the requested portion (1) during the second HTTPResponse.read(amt=N) call when the response was decompressed using the official Brotli library or (2) when HTTPResponse.drain_conn() was called after the response had been read and decompressed partially (compression algorithm did not matter here). These issues could cause urllib3 to fully decode a small amount of highly compressed data in a single operation. This could result in excessive resource consumption (high CPU usage and massive memory allocation for the decompressed data) on the client side. This vulnerability is fixed in 2.7.0.

pypdf is a free and open-source pure-python PDF library. Prior to version 6.1.3, an attacker who uses this vulnerability can craft a PDF which leads to large memory usage. This requires parsing the content stream of a page using the LZWDecode filter. This has been fixed in pypdf version 6.1.3.

A denial of service vulnerability exists in Next.js versions with Partial Prerendering (PPR) enabled when running in minimal mode. The PPR resume endpoint accepts unauthenticated POST requests with the `Next-Resume: 1` header and processes attacker-controlled postponed state data. Two closely related vulnerabilities allow an attacker to crash the server process through memory exhaustion: 1. **Unbounded request body buffering**: The server buffers the entire POST request body into memory using `Buffer.concat()` without enforcing any size limit, allowing arbitrarily large payloads to exhaust available memory. 2. **Unbounded decompression (zipbomb)**: The resume data cache is decompressed using `inflateSync()` without limiting the decompressed output size. A small compressed payload can expand to hundreds of megabytes or gigabytes, causing memory exhaustion. Both attack vectors result in a fatal V8 out-of-memory error (`FATAL ERROR: Reached heap limit Allocation failed - JavaScript heap out of memory`) causing the Node.js process to terminate. The zipbomb variant is particularly dangerous as it can bypass reverse proxy request size limits while still causing large memory allocation on the server. To be affected you must have an application running with `experimental.ppr: true` or `cacheComponents: true` configured along with the NEXT_PRIVATE_MINIMAL_MODE=1 environment variable. Strongly consider upgrading to 15.6.0-canary.61 or 16.1.5 to reduce risk and prevent availability issues in Next applications.

ProcessWire CMS 3.0.246 allows a low-privileged user with lang-edit to upload a crafted ZIP to Language Support that is auto-extracted without limits prior to validation, enabling resource-exhaustion Denial of Service.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in elixir-grpc grpc (GRPC.Compressor.Gzip, GRPC.Message modules) allows a denial of service via a gzip decompression bomb. This vulnerability is associated with program files lib/grpc/compressor/gzip.ex, lib/grpc/message.ex and program routines 'Elixir.GRPC.Compressor.Gzip':decompress/1, 'Elixir.GRPC.Message':from_data/2. 'Elixir.GRPC.Compressor.Gzip':decompress/1 calls :zlib.gunzip/1 directly on attacker-controlled bytes with no decompressed-size limit, ratio check, or incremental decoding. Because this module is the registered gzip GRPC.Compressor implementation, it is invoked automatically whenever an incoming gRPC frame carries the grpc-encoding: gzip header. :zlib.gunzip/1 allocates the entire decompressed result as a single binary, so a small highly compressible payload (for example a few kilobytes of zeros, which gzip compresses at roughly 1000:1) expands to multiple gigabytes inside a single call. The max_receive_message_length limit is enforced only against the already-decompressed message, so it provides no protection. An unauthenticated remote peer can send a single crafted frame to exhaust the BEAM node's heap and trigger an out-of-memory kill. This issue affects grpc: from 0.4.0 before 1.0.0.

cpp-httplib is a C++11 single-file header-only cross platform HTTP/HTTPS library. Prior to version 0.30.1, a Denial of Service (DoS) vulnerability exists in cpp-httplib due to the unsafe handling of compressed HTTP request bodies (Content-Encoding: gzip, br, etc.). The library validates the payload_max_length against the compressed data size received from the network, but does not limit the size of the decompressed data stored in memory.

Mattermost versions 11.4.x <= 11.4.0, 11.3.x <= 11.3.1, 11.2.x <= 11.2.3, 10.11.x <= 10.11.11 fail to validate decompressed archive entry sizes during file extraction which allows authenticated users with file upload permissions to cause a denial of service via crafted zip archives containing highly compressed entries (zip bombs) that exhaust server memory.. Mattermost Advisory ID: MMSA-2026-00598

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in elixir-tesla tesla allows a denial of service via decompression bomb in HTTP response bodies. When Tesla.Middleware.DecompressResponse or Tesla.Middleware.Compression is included in a Tesla middleware pipeline, HTTP response bodies are decompressed eagerly with no size limit. The decompress_body/2 function in lib/tesla/middleware/compression.ex passes the entire response body to :zlib.gunzip/1 or :zlib.unzip/1 without any cap on the output size. Additionally, compression_algorithms/1 splits the content-encoding header on commas and decompress_body/2 recurses once per token, applying a decompression pass on each iteration. A server advertising content-encoding: gzip, gzip, gzip, gzip causes four recursive decompression passes, yielding exponential amplification: each gzip layer can expand its input roughly 1000x, so a payload of a few hundred bytes on the wire inflates to gigabytes of BEAM heap, exhausting memory and crashing or freezing the calling process. This issue affects tesla: from 0.6.0 before 1.18.3.

IBM Concert Software 1.0.0 through 1.0.5 could allow an authenticated user to cause a denial of service due to the expansion of archive files without controlling resource consumption.

Tandoor Recipes is an application for managing recipes, planning meals, and building shopping lists. Prior to 2.6.5, a critical Denial of Service (DoS) vulnerability was in the recipe import functionality. This vulnerability allows an authenticated user to crash the server or make a significantly degrade its performance by uploading a large size ZIP file (ZIP Bomb). This vulnerability is fixed in 2.6.5.

file-type detects the file type of a file, stream, or data. From 20.0.0 to 21.3.1, a crafted ZIP file can trigger excessive memory growth during type detection in file-type when using fileTypeFromBuffer(), fileTypeFromBlob(), or fileTypeFromFile(). The ZIP inflate output limit is enforced for stream-based detection, but not for known-size inputs. As a result, a small compressed ZIP can cause file-type to inflate and process a much larger payload while probing ZIP-based formats such as OOXML. This vulnerability is fixed in 21.3.2.

AIOHTTP is an asynchronous HTTP client/server framework for asyncio and Python. Versions 3.13.2 and below allow a zip bomb to be used to execute a DoS against the AIOHTTP server. An attacker may be able to send a compressed request that when decompressed by AIOHTTP could exhaust the host's memory. This issue is fixed in version 3.13.3.

pypdf is a free and open-source pure-python PDF library. Prior to version 6.4.0, an attacker who uses this vulnerability can craft a PDF which leads to a memory usage of up to 1 GB per stream. This requires parsing the content stream of a page using the LZWDecode filter. This issue has been patched in version 6.4.0.

An issue was discovered in Cinnamon kotaemon 0.11.0. The _may_extract_zip function in the \libs\ktem\ktem\index\file\ui.py file does not check the contents of uploaded ZIP files. Although the contents are extracted into a temporary folder that is cleared before each extraction, successfully uploading a ZIP bomb could still cause the server to consume excessive resources during decompression. Moreover, if no further files are uploaded afterward, the extracted data could occupy disk space and potentially render the system unavailable. Anyone with permission to upload files can carry out this attack.

JWCrypto implements JWK, JWS, and JWE specifications using python-cryptography. Prior to 1.5.7, an unauthenticated attacker can exhaust server memory by sending crafted JWE tokens with ZIP compression. The existing patch for CVE-2024-28102 limits input token size to 250KB but does not validate the decompressed output size. An unauthenticated attacker can cause memory exhaustion on memory-constrained systems. A token under the 250KB input limit can decompress to approximately 100MB. This vulnerability is fixed in 1.5.7.

Klever-Go is the Go implementation of the Klever blockchain protocol. Prior to 1.7.17, a remote, unauthenticated denial-of-service vulnerability in Batch.Decompress (data/batch/batch.go) allows any peer that participates in a topic served by MultiDataInterceptor to allocate multi-gigabyte heaps on the receiving node from a sub-50 KiB gossip payload. A single packet is sufficient to OOM-kill a validator with conventional memory provisioning. Fleet-wide application affects chain liveness. This vulnerability is fixed in 1.7.17.

Audiobookshelf is a self-hosted audiobook and podcast server. Prior to 2.32.2, the POST /api/backups/upload endpoint decompresses the details entry from an uploaded .audiobookshelf ZIP file entirely into memory using zip.entryData(), with no limit on the decompressed size. The upload middleware also has no file size limit. An admin user can upload a crafted ZIP containing a highly compressed details entry that, when decompressed, consumes hundreds of megabytes or gigabytes of memory, crashing the server process via out-of-memory. This vulnerability is fixed in 2.32.2.

MarkUs is a web application for the submission and grading of student assignments. Prior to version 2.9.4, MarkUs currently extracts zip files without any size or entry-count limits. For example, instructors can upload a zip file to provide an assignment configuration; students can upload a zip file for an assignment submission and indicate its contents should be extracted. This issue has been patched in version 2.9.4.

PraisonAI is a multi-agent teams system. Prior to 4.5.128, the _safe_extractall() function in PraisonAI's recipe registry validates archive members against path traversal attacks but performs no checks on individual member sizes, cumulative extracted size, or member count before calling tar.extractall(). An attacker can publish a malicious recipe bundle containing highly compressible data (e.g., 10GB of zeros compressing to ~10MB) that exhausts the victim's disk when pulled via LocalRegistry.pull() or HttpRegistry.pull(). This vulnerability is fixed in 4.5.128.

In Splunk Enterprise and Universal Forwarder versions in the following table, indexing a specially crafted ZIP file using the file monitoring input can result in a crash of the application. Attempts to restart the application would result in a crash and would require manually removing the malformed file.

In python-jose 3.3.0 (specifically jwe.decrypt), a vulnerability allows an attacker to cause a Denial-of-Service (DoS) condition by crafting a malicious JSON Web Encryption (JWE) token with an exceptionally high compression ratio. When this token is processed by the server, it results in significant memory allocation and processing time during decompression.

OpenClaw versions prior to 2026.3.2 contain an archive extraction vulnerability in the tar.bz2 installer path that bypasses safety checks enforced on other archive formats. Attackers can craft malicious tar.bz2 skill archives to bypass special-entry blocking and extracted-size guardrails, causing local denial of service during skill installation.

IBM PowerVM Hypervisor FW1050.00 through FW1050.30 and FW1060.00 through FW1060.20 could allow a local user, under certain Linux processor combability mode configurations, to cause undetected data loss or errors when performing gzip compression using HW acceleration.

### Summary vLLM's `/v1/audio/transcriptions` endpoint limits compressed upload size but not decoded PCM output. A 25MB OPUS file expands to ~14.9GB of float32 PCM at decode time. Tested on vLLM v0.19.0. ### Details `SpeechToTextProcessor` rejects uploads over `VLLM_MAX_AUDIO_CLIP_FILESIZE_MB` (default 25MB) based on compressed byte length, but the audio decoder in `audio.py` accumulates all decoded frames into memory with no size limit before returning: ```python # speech_to_text.py L184-189 if len(audio_data) / 1024 ** 2 > self.max_audio_filesize_mb: raise VLLMValidationError(...) y, sr = load_audio(buf, sr=self.asr_config.sample_rate) # decoded size unchecked # audio.py L77-107 chunks: list[npt.NDArray] = [] for frame in container.decode(stream): chunks.append(frame.to_ndarray()) audio = np.concatenate(chunks, axis=-1).astype(np.float32) # single contiguous allocation ``` A 25MB OPUS file at 6kbps encodes ~8.7 hours of audio. Decoding produces ~5.7GB of float32 PCM (232x amplification), and `np.concatenate` then allocates a second contiguous array, bringing peak RSS to ~14.9GB from a single request. `SpeechToTextConfig.max_audio_clip_s` (default 30s) applies only after the full decode and does not prevent the allocation. ### Impact An unauthenticated attacker can exhaust server memory with a small number of concurrent requests, each a valid upload within the documented size limit. Severity was assessed with reference to prior OOM vulnerability reports in vLLM. ### Fix A fix for this vulnerability was merged here: https://github.com/vllm-project/vllm/pull/44970

## Impact The Compression node's Decompress operation expanded attacker-controlled archives into memory without enforcing limits on decompressed output size. An unauthenticated attacker could send a small compressed archive to a public webhook workflow using this node, causing the n8n process to terminate due to memory exhaustion and disrupting all workflows in the same instance. ## Patches The issue has been fixed in n8n version 2.24.0. Users should upgrade to this version or later to remediate the vulnerability. The fix introduces configurable limits on decompressed output size (`N8N_COMPRESSION_NODE_MAX_DECOMPRESSED_SIZE_BYTES`) and ZIP entry count (`N8N_COMPRESSION_NODE_MAX_ZIP_ENTRIES`). ## Workarounds If upgrading is not immediately possible, administrators should consider the following temporary mitigations: - Disable the Compression node by adding `n8n-nodes-base.compression` to the `NODES_EXCLUDE` environment variable. - Restrict public webhook workflows that accept archive file uploads to authenticated endpoints only. These workarounds do not fully remediate the risk and should only be used as short-term mitigation measures.

Tornado's gzip decompression routines work in limited-size chunks, but have no overall limit for the total size of decompressed chunks that they will accumulate (There has always been a limit for the total *compressed* size). This allows a malicious server to consume effectively unlimited amounts of memory if it is accessed via SimpleAsyncHTTPClient in its default configuration. `HTTPServer` is not affected in its default configuration, but it is if `decompress_request=True` is set. This bug is fixed in Tornado 6.5.6. `max_body_size` is now checked both for the compressed and cumulative decompressed size of the response. Prior to upgrading, this issue can be mitigated by setting `decompress_response=False` or using `CurlAsyncHTTPClient`.

### Summary During cleanup it is possible for a compressed request body to be decompressed into memory in one chunk. ### Impact An attacker may be able to send a compressed payload in specific situations that could be decompressed into memory, potentially leading to DoS (a zip bomb edge case). ### Workaround Disable compression if unable to upgrade. ----- Patch: https://github.com/aio-libs/aiohttp/commit/4f7480e474cccc6a8cc2c92ad3f17a31dedf8232

### Impact When `NIOHTTPRequestDecompressor` is configured with `.ratio(N)`, the decompression limit is enforced using the `Content-Length` header value from the incoming request rather than the actual number of compressed bytes received. Since `Content-Length` is attacker-controlled, a malicious client can supply an inflated value that causes the ratio check to always pass, effectively disabling the configured decompression limit. This allows an attacker to send a small, highly-compressed payload (a "gzip bomb") with a falsified `Content-Length` header to bypass the ratio-based protection entirely. The server will decompress the payload without limit, consuming unbounded memory and potentially causing denial of service. For example, a gzip payload containing highly repetitive data can achieve amplification ratios of several hundred to one. Under `.ratio(10)` such a payload should be rejected, but if the attacker sets `Content-Length` to match the decompressed size, the check evaluates `decompressed > decompressed * 10` which is always false, and the payload is accepted without error. Across repeated requests, this allows sustained memory amplification far exceeding the configured limits with no error raised. ### Relationship to CVE-2020-9840 GHSA-xhhr-p2r9-jmm7 (CVE-2020-9840) found that the `.size` limit checked compressed rather than decompressed bytes and recommended `.ratio` as a workaround. This advisory identifies a distinct flaw in the `.ratio` limit itself: it uses the attacker-supplied `Content-Length` header as the denominator rather than actual consumed compressed bytes. The two vulnerabilities are in the same decompression limit enforcement code but involve non-overlapping logic errors. Users who followed the CVE-2020-9840 workaround by switching to `.ratio(N)` are affected by this vulnerability. ### Patches Fixed in swift-nio-extras 1.34.1. The fix unifies the request and response decompressor implementations so that both accumulate actual compressed bytes received (`compressedLength += part.readableBytes`) rather than relying on any header-supplied value. ### Workarounds Use `.size(N)` instead of `.ratio(N)` if a fixed upper bound on decompressed output is acceptable for the application. The `.size` limit is not affected by this vulnerability as it does not reference `Content-Length`. ### Credits NIOExtras is grateful to @nathanielmiller23 for their reporting and assistance with the process.

### Impact The METS-GBS backend's XML parsing and the input document format detection lacked security controls, enabling: - XML External Entity (XXE) attacks to read local files or cause denial of service - Decompression bombs (zip bombs) to exhaust memory and disk space - Unbounded archive extraction consuming system resources An attacker could craft malicious METS-GBS archives that, when processed, could read sensitive files, exhaust system resources, or cause application crashes. ### Patches Fixed in version 2.91.0. The fix implements: - Secure XML parsing with `resolve_entities=False`, `load_dtd=False`, and `no_network=True` - Configurable limits: 300 MB total extraction size, 10 MB per file, 1000 member count - Cumulative size tracking across all extractions - Early termination when limits are exceeded - Secure format detection of METS-GBS tar archives with `_detect_mets_gbs()` method: maximum file size (10 MB per file), maximum member count (1000 members), and exception handling to gracefully fail when limits are exceeded ### Workarounds Avoid processing METS-GBS archives from untrusted sources. If necessary, pre-validate archives in an isolated environment with resource limits. ### References - Fix release: [v2.91.0](https://github.com/docling-project/docling/releases/tag/v2.91.0)

The LAPI router uses `gin-contrib/gzip` with `DefaultDecompressHandle` globally (`pkg/apiserver/controllers/controller.go`). This middleware decompresses incoming request bodies without enforcing a maximum decompressed size. The endpoints `/v1/watchers` or `/v1/watchers/login` require no authentication. An attacker can send small gzip-compressed JSON payloads that, when decompressed, result in hundreds of MB of valid JSON occupying server memory. Sending enough requests concurrently will cause LAPI to allocate excessive heap memory, leading the OS to forcibly terminate the process. This vulnerability is not exploitable from the network in default configurations, as LAPI only listens on the loopback interface. If developers' applications are using a multi-server setup, LAPI will be exposed in the network, in which case they are at risk if untrusted IPs can access it. ### Impact Exploiting this vulnerability will make LAPI unreachable, meaning that bouncers will not be able to fetch new decisions (but existing decisions will still be enforced) and log processors will not be able to send alerts, effectively denying the creation of new decisions. ### Workarounds If the LAPI is exposed on the network (either directly or through a reverse proxy), for example in the case of a multi-server deployment, restrict access to trusted IP addresses.