Just published.
The newest CVEs, straight from CVE.org and NVD.
-
CVE-2026-126353
Addressable is an alternative implementation to the URI implementation that is part of Ruby's standard library. From 2.3.0 until 2.9.1, URI template expressions containing several exploded variables generate ambiguous regular expressions that can divide a run of members among the variables in polynomially many ways. Template#match or Template#extract processing of an attacker-controlled URI can therefore consume a thread and cause denial of service. The flaw affects MRI Ruby before 3.2 and JRuby, TruffleRuby 40.0, and plus or number-sign expressions on MRI Ruby 3.2. MRI Ruby 3.3 and later, URI parsing, and Template#expand are unaffected. This issue is fixed in version 2.9.1.
-
CVE-2026-126354
Addressable is an alternative implementation to the URI implementation that is part of Ruby's standard library. From 2.2.5 until 2.9.1, Addressable::URI.normalize_path can resolve a path containing leading parent segments and later parent segments to a different path than an RFC 3986 resolver. URI#normalize, URI#join, URI#==, and URI#=== inherit the result. An application that authorizes the Addressable-normalized path but sends the original URI to another resolver can allow access outside the intended path allowlist. Applications that use the same normalized value for authorization and the request, or reject parent segments, are unaffected. This issue is fixed in version 2.9.1.
-
CVE-2026-126345
basic-ftp is an FTP client for Node.js. Prior to 6.2.4, the MLSD directory-listing parser in src/parseListMLSD.ts uses untrusted fact names from the FTP server as keys in the plain-object factHandlersByName dictionary without checking that the resolved property is an own handler. A fact named __proto__ therefore resolves to Object.prototype, bypasses the missing-handler guard, and is called as though it were a function, producing a TypeError. The error propagates through Client.list() and APIs that depend on it because it is not an FTPError. A malicious or compromised FTP server, or an on-path attacker on a plaintext FTP session, can make the affected operation reject, and the Node.js process terminates under the default behavior in Node.js 15 and later only if the application leaves that rejection unhandled. This issue is fixed in version 6.2.4.
-
CVE-2026-126346
basic-ftp is an FTP client for Node.js. Prior to 6.2.3, FTPContext.send() throws when a command contains carriage return, line feed, or NUL characters, but calls made from response handlers execute synchronously inside the control socket's data event without an exception boundary. In Client.login(), an attacker-influenced password reaches the PASS command after a server sends a 331 response, and in a resumed download an attacker-influenced remote path reaches the RETR command after a 350 response to REST. The resulting exception does not reject client.access() or downloadTo(), escapes ordinary promise error handling, and terminates the Node.js process, while the validation guard still prevents command injection. This issue is fixed in version 6.2.3.
-
CVE-2026-126348
Addressable is an alternative implementation to the URI implementation that is part of Ruby's standard library. From 2.0.0 until 2.9.1, the pure-Ruby Addressable::IDNA.to_unicode implementation decodes invalid Punycode A-labels instead of returning them unchanged. Labels that decode to ASCII text or fail to convert back to the same A-label can therefore make Addressable::URI#display_uri show a trusted-looking host while host, normalized_host, and the connection target retain the attacker-controlled encoded host. A trusted-name comparison using Addressable::IDNA.to_unicode output can also treat two different hosts as the same without displaying the URI. The native libidn implementation, to_ascii, and normalized_host are unaffected. This issue is fixed in version 2.9.1.
-
CVE-2026-127208
datamodel-code-generator before 0.84.0 contains a code injection vulnerability that allows attackers to achieve arbitrary code execution by supplying a malicious JSON Schema with a crafted x-python-type annotation containing a string ast.Constant node in a subscript slice position, which bypasses validation in parse_python_type_annotation and _python_type_expr_from_ast. Attackers can embed malicious Python expressions such as __import__ calls within quoted string literals in the generated source code, which Pydantic v2 evaluates as a ForwardRef via eval() during model_rebuild(), resulting in arbitrary code execution on developer workstations, CI/CD pipelines, or applications that generate and import the produced models at runtime.
-
CVE-2026-126205
Loofah is a general library for manipulating and transforming HTML/XML documents and fragments, built on top of Nokogiri. Prior to 2.25.3, the Loofah HTML5 sanitizer checks href and xlink:href on elements in SVG_ALLOW_LOCAL_HREF with a regular expression whose line anchor and newline-matching whitespace cause an all-whitespace value to be rescanned from every line start. An attacker can submit an SVG reference element with a long multiline-whitespace attribute and consume CPU quadratically during sanitization. CRuby 3.1 and earlier and JRuby are affected, while CRuby 3.2 and later uses linear-time regular-expression matching. The default safelist reaches the vulnerable check without a configuration change. This issue is fixed in version 2.25.3.
-
CVE-2026-111456
A missing authorization check has been found in ConsumerResource.setupOwner() function of consumer registration module in the Candlepin package. The registration method checks permission to create a consumer only when the caller is a user. A consumer that authenticates with its identity certificate does not go through that check. A low-privileged network attacker or user who has a valid consumer certificate for one organization can register a new consumer in a different organization by supplying that organization's key. The response includes an identity certificate for the new consumer. With that certificate, the attacker can view the target organization's subscription pools. When the organization is not using simple content access, the attacker can attach that organization's entitlements to the new consumer.
-
CVE-2026-107152
An improper authentication vulnerability has been found in the task console in the smart_proxy_dynflow package. The console accepts any client certificate signed by the capsule's certificate authority. It does not check the list of servers the capsule trusts. A low-privileged network attacker or user must already hold such a certificate. They can open the console and read job details. Those details include passwords when a job was given them. They can also cancel a running task.
-
CVE-2026-107360
A flaw was found in rubygem-safemode. Compound assignment expressions in sandboxed code are evaluated without mediating method calls through the Jail allow-list. An authenticated attacker who can supply template content can invoke arbitrary no-argument methods on objects exposed to the template renderer without Jail mediation, including destructive operations such as host deletion and disclosure of non-allow-listed object state.
-
CVE-2026-18379
A flaw was found in yggdrasil-worker-package-manager. An authenticated attacker can exploit this issue to achieve arbitrary file modification or potential privilege escalation. The vulnerability exists because repository management operations fail to sanitize directory traversal sequences (such as '../') within repository names. By supplying a crafted repository name, an attacker can manipulate the root-level worker process into creating or deleting files outside the intended directory, compromising host system integrity.
-
CVE-2026-127119
Unverified ownership of the default cluster log location in the EMR module in AWS SDK for pandas (awswrangler) from 1.1.0 before 3.18.0 might allow remote actors to read Amazon EMR cluster logs, including job step output, via a pre-created S3 bucket. To remediate this issue, users should upgrade to version 3.18.0 or later. Users who can't upgrade immediately should pass an explicit log location, or create the default EMR log bucket in their own account in each region they use.
-
CVE-2026-127030
A flaw was found in freeradius. An improper length validation in the EAP-FAST module allows a stack-based buffer overflow during session ticket processing. A remote, unauthenticated attacker can exploit this flaw by sending a specially crafted handshake request, causing the server process to crash and resulting in a Denial of Service (DoS).
-
CVE-2026-127029
A flaw was found in freeradius's Extensible Authentication Protocol Method for Subscriber Identity Module (EAP-SIM) module. When handling authentication requests, the server copies a client-supplied identity into an internal structure without properly checking its length against a fixed-size buffer. An unauthenticated remote attacker can exploit this issue by sending specially crafted EAP-SIM messages, triggering a stack buffer overflow that crashes the service and results in a Denial of Service (DoS).
-
CVE-2026-126835
The MongoDB Node.js driver does not consistently enforce its permitted-host restriction during OIDC authentication. A bearer token produced by the embedding application's own login callback may therefore be transmitted to a host outside the configured set of permitted hosts, which may result in unintended disclosure of that credential. No database account is required; the party supplying the connection configuration needs only to influence the settings the application passes to the driver.
-
CVE-2026-126190
Soup Sieve is a CSS selector library designed to be used with Beautiful Soup 4. Prior to 4.0, Soup Sieve repeatedly walks the same ancestors while matching descendant combinators, and each additional descendant combinator raises the polynomial degree of the work on deeply nested attacker-influenced HTML or XML. Applications using selectors such as section div div through select(), select_one(), or Beautiful Soup CSS APIs against untrusted documents can have a worker occupied for seconds or minutes, while trusted-only document processing is not affected. Selectors compiled with NOCACHE continue to use the old uncached behavior. This issue is fixed in version 4.0.
-
CVE-2026-126192
Soup Sieve is a CSS selector library designed to be used with Beautiful Soup 4. Prior to 2.10, matching a developer-authored selector that places the general sibling combinator inside :has() repeatedly scans each candidate element's following siblings when the right-hand side is absent, producing quadratic work on attacker-controlled HTML. Applications that evaluate a fixed selector such as div:has(a ~ b) over untrusted markup can have a worker occupied for minutes by a small document, while the adjacent sibling form and trusted-only document processing are not affected in the same way. This issue is fixed in version 2.10.
-
CVE-2026-127109
A code injection vulnerability in promzard’s template-loading functionality allows an attacker who can control property names in the context object (ctx) to execute arbitrary code with the privileges of the Node.js process. The library inserts context property names into a generated function’s parameter list without validation or escaping and evaluates the resulting source using `vm.runInThisContext()`. A crafted property name can escape the parameter list and inject JavaScript code, even when the template file is fully trusted.
-
CVE-2026-97845
An uncontrolled recursion vulnerability has been discovered in the Qt QXmlSimpleReader XML parsing functionality of the Qt 5 Core Compatibility APIs module (Qt 6) and the Qt XML module (Qt 5).
-
CVE-2026-127056
PapaParse before 5.7.1 contains an inefficient algorithmic complexity vulnerability that allows remote unauthenticated attackers to cause denial of service by submitting delimiter-free quoted CSV input to Papa.parse(). Attackers can supply a single-column quoted CSV with no delimiters, triggering repeated full-buffer indexOf scans per row that grow quadratically with input size, exhausting CPU and blocking the Node.js event loop or freezing a browser tab.
-
CVE-2026-126994
A vulnerability has been found in braindecode up to 1.8.1. Affected is the function resolve_type_kwargs of the file braindecode/models/util.py of the component Model Config Deserialization. Such manipulation of the argument Activation leads to use of externally-controlled input to select classes or code. The attack can be launched remotely. The exploit has been disclosed to the public and may be used. Upgrading to version 2.0.0 is able to address this issue. The name of the patch is 05b764d93b9a50400f523206fbe03d73d0cf5edb. The affected component should be upgraded.
-
CVE-2026-114369
Resource exhaustion vulnerability in the Qt Framework QWebSocket and QWebSocketServer classes of the Qt WebSockets module
-
CVE-2026-126644
A link-following flaw was found in the authorized_key module of the ansible.posix collection, residual to the fix for CVE-2026-11837. That fix guards only the final path component (os.chown with follow_symlinks=False and O_NOFOLLOW in _safe_open_write()). When an unprivileged local user replaces their ~/.ssh directory with a symlink to a directory they do not own, and an operator subsequently runs the authorized_key task as root (a common provisioning pattern), the module resolves its path operations through the symlink: atomic_move() writes the temporary file in the symlink target and renames it to <target>/authorized_keys, creating or overwriting a file in the attacker-chosen directory, and os.chmod(sshdir, 0o700) follows the symlink and changes the target directory's permissions. The written file is chowned to the unprivileged user, so this is not a privilege escalation, but it allows an attacker to create or overwrite files (for example in /etc/apt/apt.conf.d/, which APT reads regardless of file ownership) to cause integrity and availability impacts.
-
CVE-2026-82292
A NULL pointer dereference vulnerability in the Qt Framework QDomDocument::setContent() XML parsing path of the Qt XML module (qtbase).
-
CVE-2026-12507
Zohocorp ManageEngine AssetExplorer versions below 7860, ServiceDesk Plus versions below 15290, ServiceDesk Plus MSP versions below 15150, and SupportCenter Plus versions below 15150 are affected by a authenticated remote code execution vulnerability in the Reports module. An authenticated user with the Create Query Reports permission could exploit specially crafted database queries to execute arbitrary commands on the underlying server.
-
CVE-2026-16550
An external control of file name or path vulnerability in the Media Parser in Synology DiskStation Manager (DSM) before 7.3.2-86009-5 and 7.4.1-90080-1 allows remote authenticated users to write arbitrary files.
-
CVE-2025-10935
libilbm ILBM File Parsing Heap-based Buffer Overflow Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of libilbm. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file. The specific flaw exists within the parsing of ILBM files. The issue results from the lack of proper validation of the length of user-supplied data prior to copying it to a heap-based buffer. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-27792.
-
CVE-2026-126785
A vulnerability has been found in rmalright TR-VISION HOME 2.0.5. Impacted is the function sub_11504 in the library WFanManager.sys of the component IOCTL Handler. Such manipulation leads to incorrect default permissions. The attack must be carried out locally. The vendor was contacted early about this disclosure but did not respond in any way.
-
CVE-2026-126673
Data::Dump::Streamer versions before 3.00 for Perl allow code injection via unescaped Regexp patterns in _dump_qr. A Regexp's pattern is written into a qr literal without escaping "$" or "@", so "@{[ ... ]}" and "${\ ... }" in the pattern run as code when the dump is evaluated. An already escaped delimiter gets a second backslash, which leaves the delimiter bare and ends the qr literal early. An attacker who supplies a pattern that an application compiles, dumps and later restores with eval or do can run arbitrary Perl code.
-
CVE-2026-65950
A path traversal vulnerability in the ZIP extraction logic of Lenovo XClarity Essentials OneCLI 5.5.0 for Linux could allow an attacker to write arbitrary files with the privileges of the OneCLI process by convincing a user to process a specially crafted ZIP-based update package.
-
CVE-2026-65951
A path traversal vulnerability in the TAR.GZ extraction logic of Lenovo XClarity Essentials OneCLI 5.5.0 for Linux could allow an attacker to write arbitrary files with the privileges of the OneCLI process by convincing a user to process a specially crafted TAR.GZ update package.
-
CVE-2026-65982
nginx-http-flv-module is a media streaming server based on nginx-rtmp-module. Prior to version 1.2.14, if a server received an HTTP/1.0 request without a header Host, which was permitted by HTTP core of Nginx, then r->headers_in.host was NULL, nginx-http-flv-module used r->headers_in.host->value to make a pseudo tcUrl, so NULL-pointer dereference happened. This issue has been patched in version 1.2.14.
-
CVE-2026-65833
iopress is a native HTTP server for Node.js with platform-specific async I/O. Prior to version 1.0.5, the chunked-transfer-encoding HTTP request parser in @iopress/core reads past the end of the input buffer when a request declares a chunk size larger than the body actually received. This results in a heap-buffer-over-read. This issue has been patched in version 1.0.5.
-
CVE-2026-123871
Node.js discloses proxy credentials in ERR_PROXY_INVALID_CONFIG error messages. CVE-2026-48615 established that a proxy credential must not appear in a Node.js error message and redacted the username and password from ProxyConfig.href, but that fix missed four error paths in the same subsystem, one of them inside the same constructor, which interpolate the raw proxy URL, credentials included, into an ERR_PROXY_INVALID_CONFIG message. The proxy URL comes from the environment, but the defect is the sink: a secret that Node.js redacts everywhere else is emitted verbatim into a much wider audience than the process environment, such as stderr, container logs, log aggregation, and crash reporting services. This vulnerability affects Node.js 26.x, 24.x, and 22.x.
-
CVE-2026-123864
A re-entrancy flaw in Node.js HTTPS lets an unauthenticated client read memory belonging to other connections, including TLS private keys and other users' requests. TLSWrap::DoWrite() calls ClearOut() directly when asked to write zero bytes; if that write is issued from JavaScript running inside the HTTP parser's on_body callback, ClearOut() decrypts the next TLS record while the parser is still mid-body, re-entering llhttp_execute() and causing a use-after-free and an out-of-bounds heap read. The server then copies that freed or unrelated heap memory into the response for the request body. Triggering it requires a handler that calls res.flushHeaders() before the request body ends and later calls res.end() with a non-empty chunk, and reading the leaked memory requires any ordinary route that echoes or stores a request body. Cleartext node:http is not affected. This vulnerability affects Node.js 26.x, 24.x, and 22.x.
-
CVE-2026-91872
Dell Ansible Storage for PowerStore, versions prior to 3.9.0,contains an Insertion of Sensitive Information into Log File vulnerability in the configuration module. A low privileged attacker with local access could potentially exploit this vulnerability, leading to information exposure.
-
CVE-2026-65820
aeon is a scikit-learn compatible Python library for learning from time series. Prior to version 1.5.0, a GitHub Actions workflow used pull_request_target for pre-commit checks on pull requests from forks. In this configuration, fork-controlled code could be checked out and executed in a privileged context. The vulnerable workflow also generated/used elevated credentials and could automatically commit or push fixes back to pull request branches. This created a potential token exposure and repository write path if an attacker submitted a malicious fork pull request designed to execute during the privileged workflow run. This issue has been patched in version 1.5.0.
-
CVE-2026-64855
soapbar is a SOAP library for Python — client, server, and WSDL handling. Prior to version 0.13.0, verify_envelope_bsp() derived its trust anchor from the message it was verifying. It read the X.509 certificate out of the caller-untrusted wsse:BinarySecurityToken, validated only that certificate's not-before / not-after dates, and then verified the signature against that same certificate. There was no CA-chain validation, no issuer/subject allowlist, and no pinning. A successful verify_envelope_bsp() therefore proved only that the envelope was signed by whoever supplied the embedded certificate — never that it was signed by a trusted party. An attacker who mints their own certificate, signs a forged envelope, and embeds that certificate in the token passes verification. Any consumer that treats a successful BSP verification as authentication can be impersonated (signature forgery / authentication bypass). verify_envelope() (which pins a caller-supplied x509_cert) was not affected — the flaw was specific to the BSP variant whose trust anchor came from the message. This issue has been patched in version 0.13.0.
-
CVE-2026-115819
Calling the wide scanf family of functions (swscanf, fwscanf, wscanf, vswscanf, vfwscanf and vwscanf) with a %s or %[ conversion without the 'l' length modifier in the GNU C Library version 2.2 to version 2.44 writes an extra null byte one past the end of the destination array when the converted multibyte string and its terminating null byte exactly fill that array. These conversions store each wide input character into the caller's char array as a multibyte sequence. The conversion ends with the sequence that returns to the initial shift state, which also stores the terminating null byte, and the functions then append a second null byte. A destination sized for the converted string and one terminator, as the C standard requires the caller to provide, is therefore written one byte past its end. The out-of-bounds byte is always a null byte, which may corrupt adjacent data or crash the process. An attacker who supplies the input can trigger the overflow with a string whose converted length exactly fills the destination. The narrow scanf functions, the %ls and %l[ conversions into a wchar_t array, and the 'a' and 'm' allocation modifiers are not affected, and storing a wide scanf conversion into a narrow char array is uncommon. This issue affects all architectures and ABIs supported by the GNU C Library.
-
CVE-2026-100290
The regcomp and regexec functions in the GNU C Library version 2.3 to version 2.44 may read freed memory and free the same heap block twice when a memory allocation fails while a regular expression is compiled or matched. The re_dfa_add_node (the NFA node table) and build_range_exp (bracket ranges) grow several arrays with consecutive realloc calls and store the new pointers only after all of them succeed. When a later call fails, the pattern buffer keeps pointers to blocks that were already freed, and the error cleanup frees them again. In regexec, the back-reference cache (match_ctx_add_entry) frees its array when a realloc fails but keeps the pointer, which is freed again on the error path. An attacker does not need to supply a crafted regular expression: patterns large enough to grow these arrays are common. The failure has to hit a specific allocation, though, so the issue can only be reached when the system is already under memory pressure. An attacker who supplies the pattern or the subject string has little control over which allocation fails, so the expected impact is a crash of the process. This issue affects all architectures and ABIs supported by the GNU C Library.
-
CVE-2026-76981
Zoho Corporation ManageEngine ADAudit Plus versions below 8711 are vulnerable to an authenticated second order SQL injection vulnerability affecting ADAudit Plus custom reports module.
-
CVE-2026-97038
Building a malicious package containing ".syso" files or cgo linker inputs could inject arbitrary linker flags or execute arbitrary code at link time. Linker input files—including ".syso" files, direct object or library inputs in "#cgo LDFLAGS", and libraries in "-L" directories—were passed to the external linker without validating their format or checking for embedded linker options. Across platforms, this allowed executing GNU ld implicit linker scripts disguised as libraries or ".syso" files (or loaded via "-l:<filename>" or "-l" containing path separators), injecting unallowlisted linker flags via PE/COFF ".drectve" sections on Windows or Mach-O "LC_LINKER_OPTION" load commands on Darwin, or invoking LTO plugins via GCC or LLVM LTO intermediate representation sections in ELF, PE/COFF, or Mach-O objects. The go command now verifies that ".syso" files, direct linker inputs, and libraries in "-L" directories are valid binary objects or archives that do not contain unallowlisted ".drectve" directives, "LC_LINKER_OPTION" load commands, or LTO IR sections, and disallows "-l:<filename>" and ":", "=", or path separators in "-l" flags.
-
CVE-2026-97039
Building a malicious package that uses "#cgo pkg-config:" directives could load untrusted pkg-config ".pc" files or inject unvalidated compiler and linker flags. The pkg-config flag allowlist previously permitted "--with-path=" and "--personality=", allowed "--define-variable=" values beginning with "=" or "$" (which can rewrite paths in system ".pc" files into sysroot-prefixed "-L=" or "-L$SYSROOT" linker search paths), and allowed package arguments ending in ".pc" or containing path separators or "..", which caused pkg-config to load local ".pc" files directly from the package directory. In addition, flags returned by "pkg-config --cflags" in cgoCPPFLAGS were not checked when determining whether internal linking is safe. The go command now disallows "--with-path=", "--personality=", leading "=" or "$" characters in "--define-variable=" values, and local ".pc" paths in "#cgo pkg-config:" directives, and inspects cgoCPPFLAGS when checking internal linking compatibility.
-
CVE-2026-97040
Building a package containing a crafted ".syso" file could bypass cgo flag validation and inject arbitrary linker flags when linking. When loading ".syso" entries from a package archive, the linker only checked for recognized host object headers (ELF, Mach-O, PE, and XCOFF) before falling through to Go object parsing. A ".syso" file beginning with a "go object" header was parsed as a Go compiler object, allowing untrusted cgo metadata directives (such as "cgo_ldflag" or "cgo_dynamic_linker") embedded in the ".syso" file to be passed to the external linker without validation. Additionally, ".syso" filenames longer than 16 bytes lost their ".syso" extension when truncated in archive headers, and unrecognized ".syso" formats forced external linking in pure-Go builds. The linker now rejects ".syso" files that are not recognized host object formats, and "cmd/go" and "cmd/pack" preserve the ".syso" suffix when truncating long archive member names.
-
CVE-2026-97041
Building a malicious package containing "//go:cgo_*" compiler directives could inject arbitrary flags into the external linker invocation on ELF platforms. The compiler and linker did not reject commas in "//go:cgo_import_dynamic", "//go:cgo_export_dynamic", "//go:cgo_export_static", "//go:cgo_import_static", and "//go:cgo_dynamic_linker" directives, and the linker passed the dynamic linker path using "-Wl,--dynamic-linker,<path>". Furthermore, the two-argument form of "//go:cgo_import_dynamic" (which is permitted in non-cgo Go source files) could overwrite the external symbol name of a symbol already marked for cgo export. A symbol name or dynamic linker path containing a comma or leading "@" could split "-Wl,--export-dynamic-symbol=" or "-Wl,--dynamic-linker," into arbitrary host linker flags or response file arguments. The compiler and linker now reject commas in "//go:cgo_*" symbol and dynamic linker directives, the linker passes "-Wl,--dynamic-linker=<path>", and "cgo_import_dynamic" no longer overwrites symbols exported via "cgo_export_dynamic" or "cgo_export_static".
-
CVE-2026-97042
Building a malicious package that uses "//go:cgo_import_dynamic" or links against a crafted PE object on Windows could pass unsafe library arguments to the external linker. The compiler only checked the first character of the library name in "//go:cgo_import_dynamic" directives and skipped validation in cgo-generated files. In addition, on Windows, "cmd/cgo" and "cmd/link" did not validate imported DLL names extracted from PE object files before emitting "-l<name>" flags for the external linker, allowing a crafted DLL name containing ".." path traversal or special characters to reference arbitrary files on disk. The compiler now validates every character of the library name in "//go:cgo_import_dynamic" directives (including in cgo-generated files on Windows) and rejects ".." path traversal, and "cmd/cgo" and "cmd/link" reject PE imported DLL names containing path separators, "..", ":", or leading "-" or "@".
-
CVE-2026-97043
Building a malicious package that uses cgo could bypass "#cgo LDFLAGS" validation and cause unexpected code execution at build time, or cause the go command to panic, via crafted quoting in "//go:cgo_ldflag" directives. When "cmd/cgo" generates "_cgo_gotypes.go", it quotes each "#cgo LDFLAGS" argument with strconv.Quote in a "//go:cgo_ldflag" directive, but "cmd/compile" and "cmd/go" parsed those directives by splitting on raw quote characters without handling backslash escapes or unquoting the string. This mismatch allowed disallowed linker flags or unvalidated linker search paths to be smuggled past the go command's flag checks and passed to the external linker. In addition, deduplicating LDFLAGS in the go command could panic with an out-of-bounds slice index when a prefix of the flag list matched near the end of the slice. "cmd/compile" and "cmd/go" now properly handle backslash-escaped quotes and unquote "//go:cgo_*" pragma strings with strconv.Unquote, and the slice bounds check in processCgoOutputs has been fixed.
-
CVE-2026-97054
For HTTP/1 servers, when MaxBytesReader's limit is exceeded, "Connection: close" header is automatically set without synchronization, which might race with another header being set explicitly in the server handler. As a result, an HTTP/1 server that utilizes MaxBytesReader or MaxBytesHandler can crash due to runtime panic if its handler reads the request body concurrently with writing the response header.
-
CVE-2026-97033
Building a malicious package that uses cgo could cause arbitrary code execution at build time through several compiler and linker flags permitted in "#cgo" directives. Specifically, "--sysroot" in GCC and Clang relocates compiler subprograms, spec files, and CRT startup objects to an attacker-controlled directory; "-target" in Clang falls back to executing target helper tools from the current working directory when not found in PATH; "-isysroot" and "-Wl,-syslibroot" can pass a regular file on disk to GNU ld.bfd, which interprets it as a positional linker script; and "-x" permitted specifying arbitrary compiler input languages beyond those supported by cgo. "--sysroot" and "-target" are now disallowed in "#cgo" flags, "-isysroot" and "-Wl,-syslibroot" are verified not to point to regular files, and "-x" is restricted to the C, C++, Objective-C, and assembler languages supported by cgo.
-
CVE-2026-97034
When using GCC, building a malicious package that uses cgo could cause arbitrary command execution at link time via untrusted GCC spec files ("specs" or "*.spec"). When linking with "-fopenmp", "-fopenacc", or "-fsanitize=" in "#cgo LDFLAGS", GCC's built-in link spec evaluates "%:include(libgomp.spec)" or "%:include(libsanitizer.spec)" and searches for spec files in "-L" and sysroot directories (including subdirectories and multilib sibling directories such as "../lib64"), as well as the working directory when "-fopenmp" or "-fopenacc" is used or when "-fsanitize=" is combined with "-static" or "-static-lib*". The go command now rejects "-L", "-isysroot", and "-Wl,-syslibroot" directories (and the working and package directories when applicable) that contain GCC spec files, and disallows "@" and "-" prefixes in "-fsanitize=" compiler flag values.
The newest published CVE records from CVE.org and NVD, refreshed every few minutes, with severity, exploit odds and fix status.