Pillow versions prior to 13.0.0 contain a vulnerability in the TarIO initializer where a negative octal member size causes an infinite CPU spin loop. This occurs because the code computes a backward seek that returns to the same 512-byte header repeatedly when the requested member name does not match, leading to high CPU consumption. This issue only affects applications that directly use PIL.TarIO.TarIO instead of the default Image.open() method. The vulnerability is fixed in Pillow version 13.0.0.
Pillow versions prior to 13.0.0 contain a vulnerability in the PILfont text-width accumulation logic where negative glyph advances are not properly checked, leading to signed integer overflow. This can cause logic errors or excessive resource consumption in applications using text size or layout calculations. The issue is fixed in version 13.0.0.
A vulnerability in the Pillow Python imaging library versions 5.1.0 through 12.x allows a specially crafted PDF with malformed shared page-tree structures to cause resource exhaustion during PDF append operations. This occurs because the library treats the PDF /Pages graph as a tree without detecting duplicate references, leading to exponential expansion of page references and excessive CPU consumption. The issue is fixed in version 13.0.0.
Pillow versions from 11.0.0 up to but not including 13.0.0 contain a vulnerability in multi-frame image loaders where stale native image storage is reused after mode transitions. This can cause out-of-bounds writes and crashes, and in some cases, native over-reads leading to potential command execution. The issue is fixed in version 13.0.0.
A vulnerability in the Pillow Python imaging library versions 9.1.0 through 12.x allows crafted BMP RLE images to cause excessive memory allocation during decoding. This memory amplification can lead to denial of service by exhausting system memory. The issue is fixed in version 13.0.0.
Pillow versions prior to 13.0.0 contain a vulnerability in SpiderImagePlugin.seek() where the decompression bomb check is bypassed. This allows an attacker to craft a file that initially appears safe but later declares very large image dimensions, causing unbounded memory allocation and potential denial of service. The issue is fixed in version 13.0.0.
Pillow versions prior to 13.0.0 contain a vulnerability where the decompression-bomb protection can be bypassed via multi-frame seek() operations in MPO and DCX image files. This allows a crafted multi-frame image to pass initial checks with a small first frame but later declare an excessively large frame, leading to memory exhaustion and denial of service when that frame is decoded. The issue is fixed in version 13.0.0.
Pillow versions from 11.2.1 up to but not including 13.0.0 have a vulnerability in the Arrow export feature where stale heap memory can be leaked due to incorrect handling of image row alignment. This can cause uninitialized or previously freed memory to be included in the exported Arrow buffer, potentially disclosing sensitive data between users. The issue is fixed in version 13.0.0.
A buffer overflow vulnerability exists in the Brotli decoder used by Google Brotli versions 1.0.3 through 1.2.0, affecting PowerShell versions 7.4.0 through 7.4.21, 7.5.0 through 7.5.12, and 7.6.0 through 7.6.8. This flaw allows a remote unauthenticated attacker to corrupt adjacent memory, potentially leading to sensitive data disclosure, denial of service, or arbitrary code execution. Fixed versions are 7.4.22, 7.5.13, and 7.6.9.
A flaw in Node.js crypto affects the internal entropy caches used by crypto.randomUUID(), crypto.randomUUIDv7(), and crypto.randomInt() when user-land startup snapshots or single executable applications are built. If these functions are called during snapshot creation, processes started from that snapshot return identical UUIDs and random integers until the cache is refilled, causing collisions and predictability across processes. This impacts Node.js versions 22.x, 24.x, and 26.x prior to specific fixed releases.
A flaw in the Node.js Permission Model filesystem enforcement causes fs.realpathSync() to bypass required read permission checks. This incomplete fix of CVE-2026-21715 allows code running with restricted --allow-fs-read permissions to use fs.realpathSync() to access filesystem paths outside permitted directories. The vulnerability affects Node.js versions 22.x, 24.x, and 26.x prior to specific fixed releases.
A vulnerability in Node.js TLS hostname verification allows a certificate with only a DNS Subject Alternative Name (SAN) containing a numeric IP address to authenticate an IP endpoint when the client connects using non-canonical IPv4 address formats. This occurs because the function tls.checkServerIdentity() normalizes non-canonical IP addresses to their canonical form and matches them against DNS names in the certificate, incorrectly accepting certificates with DNS SANs for numeric IPs. The flaw affects Node.js versions 22.x, 24.x, and 26.x prior to specific fixed releases.
A flaw in the Node.js Permission Model for filesystem enforcement allows access to filesystem paths that do not exactly match a wildcard permission rule. Specifically, when a requested path is shorter than the prefix of a wildcard grant, the permission check incorrectly treats it as a match, potentially granting unintended access. This affects Node.js versions 25.0.0 up to but not including 26.12.1 and is fixed in version 26.12.1.
A vulnerability in the Node.js zlib.ZipFile API allows it to return bytes not belonging to the requested ZIP entry when local and central header validation fails. This occurs because the parsed local header is cached without a content offset, causing subsequent reads to return attacker-controlled bytes from the start of the file. The CRC-32 check still passes due to attacker control of the central directory CRC. This can cause Node.js-based ZIP scanners to approve different content than other ZIP implementations extract. The flaw affects Node.js versions 25.0.0 up to but not including 26.12.1 and is fixed in 26.12.1.
A vulnerability in Node.js TLS hostname verification allows a server certificate to be validated against a different IP address than the one the client connected to when the host is specified as a non-canonical IPv4 literal with a leading zero. This occurs because Node.js normalizes the host using WHATWG URL rules interpreting leading-zero octets as octal, while some operating systems interpret them as decimal. This mismatch can cause a server at one IP to be authenticated with a certificate issued for another IP. The flaw affects Node.js versions 22.x, 24.x, and 26.x prior to their respective fixed releases.
A vulnerability in Node.js TLS hostname verification allows a certificate without a matching IP address Subject Alternative Name (SAN) to authenticate a connection to an IPv4 address when the host is represented with non-ASCII characters such as fullwidth digits. This occurs because the tls.checkServerIdentity() function fails to recognize such hosts as IP addresses and instead matches them against DNS names and Common Name fields in the certificate. This flaw affects Node.js versions 24.x and 26.x and was introduced as a regression after a previous fix. Fixed versions are 24.22.1 and 26.12.1.
A vulnerability in Node.js DNS resolution functions allows hostnames containing a NUL byte to be silently truncated by dns.resolve*() and dns.reverse(). This can cause applications that validate hostnames but then resolve them using these functions to act on DNS records for a different hostname than intended. The flaw affects Node.js versions 22.x, 24.x, and 26.x prior to specific fixed releases.
A flaw in Node.js HTTP Agent connection reuse allows a client to accept a response sent before the client has sent the request. This occurs because the guard that destroys idle keep-alive sockets on unexpected data only runs when the event loop polls for I/O, and unsolicited responses may be misinterpreted as valid responses to new requests. This vulnerability affects Node.js versions 22.x, 24.x, and 26.x and is a partial regression of a previous flaw (CVE-2026-48931).
OpenBao versions from 2.5.0 up to but not including 2.6.2 contain a vulnerability in the PKI role allowed_ip_sans_cidr enforcement. When use_csr_sans is enabled, authenticated users with sign permission can submit a CSR containing IP SANs outside the allowed range and obtain CA-signed certificates for unauthorized IP addresses. This issue is fixed in version 2.6.2.
OpenBao versions from 2.3.0 up to but not including 2.6.3 have an improper authorization vulnerability in the plugin catalog APIs. This flaw allows privileged child-namespace tokens to modify plugin registrations because the /v1/sys/plugins/catalog/* endpoint is not restricted to the root namespace. Exploitation can lead to removal of existing plugin registrations, creation of invalid registrations, or unauthorized registration of installed plugins via a child namespace's sys/ mount, resulting in denial of service.