Buffer Overflow
A buffer overflow occurs when a program writes more data to a memory buffer than it was allocated to hold, causing the excess data to spill into adjacent memory regions.
How It Works
A buffer overflow occurs when a program writes more data to a memory buffer than it was allocated to hold, causing the excess data to spill into adjacent memory regions. This overwrites whatever data or code exists there, corrupting program state and potentially giving attackers control over execution flow.
Stack-based overflows are the most common variant. When a function allocates a fixed-size buffer on the stack and then copies user-controlled input without proper bounds checking, attackers can overflow past the buffer to overwrite the function's return address. When the function completes, instead of returning to legitimate code, execution jumps to attacker-specified memory containing malicious shellcode. Heap-based overflows work differently—they corrupt heap metadata like chunk size fields or free list pointers, leading to arbitrary memory writes when the allocator processes the corrupted structures.
Modern exploitation bypasses defensive mechanisms through techniques like Return-Oriented Programming (ROP), which chains together existing code snippets to avoid non-executable memory protections. Attackers may also use heap spraying to reliably position shellcode at predictable addresses, defeating address randomization.
Impact
- Remote code execution — attacker gains ability to run arbitrary commands with the privileges of the vulnerable process
- Privilege escalation — exploiting kernel or setuid program overflows to gain root/SYSTEM access
- Denial of service — crashes and memory corruption that render systems unusable
- Information disclosure — reading sensitive data from adjacent memory regions that should be inaccessible
- Authentication bypass — overwriting security-critical variables like permission flags or user IDs
Real-World Examples
Fortinet FortiOS suffered a critical buffer overflow (CVE-2025-32756) that allowed unauthenticated remote attackers to execute code as root on firewalls and VPN gateways. Attackers actively exploited this to compromise enterprise network perimeters before patches were available.
The Slammer worm from 2003 exploited a stack overflow in Microsoft SQL Server, spreading to 75,000 hosts in ten minutes by sending a single malformed UDP packet that overwrote the return address with shellcode. No authentication was required.
OpenSSH historically contained a heap overflow in challenge-response authentication that allowed pre-authentication remote root compromise on Unix systems, demonstrating how memory corruption in privileged network services creates maximum impact scenarios.
Mitigation
- Memory-safe languages — Rust, Go, and modern managed languages prevent buffer overflows by design through automatic bounds checking
- Stack canaries — random values placed before return addresses that detect corruption before control transfer
- Address Space Layout Randomization (ASLR) — randomizes memory locations making exploitation less reliable
- Data Execution Prevention (DEP/NX) — marks memory regions as non-executable, preventing direct shellcode execution
- Bounds checking — validate input sizes before copying, use safe functions like
strncpyinstead ofstrcpy - Fuzzing and static analysis — automated testing to discover overflows before deployment
Recent CVEs (36676)
In libass 0.14.0, the `ass_outline_construct`'s call to `outline_stroke` causes a signed integer overflow. Rated high severity (CVSS 8.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available and no vendor patch available.
An issue was discovered in QEMU through 5.1.0. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
An exploitable code execution vulnerability exists in the fsck_chk_orphan_node functionality of F2fs-Tools F2fs.Fsck 1.13. Rated high severity (CVSS 7.8), this vulnerability is no authentication required, low attack complexity. Public exploit code available and no vendor patch available.
An exploitable information disclosure vulnerability exists in the get_dnode_of_data functionality of F2fs-Tools F2fs.Fsck 1.13. Rated medium severity (CVSS 5.5), this vulnerability is no authentication required, low attack complexity. Public exploit code available and no vendor patch available.
Heap-based buffer overflow in archive_string_append_from_wcs() (archive_string.c) in libarchive-3.4.1dev allows remote attackers to cause a denial of service (out-of-bounds write in heap memory. Rated medium severity (CVSS 6.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity.
In appendFormatV of String8.cpp, there is a possible out of bounds write due to incorrect error handling. Rated high severity (CVSS 7.8), this vulnerability is low attack complexity.
In setUpdatableDriverPath of GpuService.cpp, there is a possible memory corruption due to a missing permission check. Rated high severity (CVSS 7.8), this vulnerability is low attack complexity.
In gatt_process_read_by_type_rsp of gatt_cl.cc, there is a possible out of bounds read due to a missing bounds check. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. This Out-of-bounds Read vulnerability could allow attackers to read data from memory outside the intended buffer boundaries.
In ~AACExtractor() of AACExtractor.cpp, there is a possible out of bounds write due to uninitialized data. Rated medium severity (CVSS 6.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. This Out-of-bounds Write vulnerability could allow attackers to write data beyond allocated buffer boundaries leading to code execution or crashes.
In remove of String16.cpp, there is a possible out of bounds write due to an integer overflow. Rated high severity (CVSS 7.8), this vulnerability is low attack complexity. No vendor patch available.
In gatt_process_read_by_type_rsp of gatt_cl.cc, there is a possible out of bounds read due to a missing bounds check. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
There is a possible out of bounds read due to a missing bounds check.Product: AndroidVersions: Android SoCAndroid ID: A-163003156. Rated critical severity (CVSS 9.1), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
There is a possible out of bounds read due to a missing bounds check.Product: AndroidVersions: Android SoCAndroid ID: A-163008256. Rated critical severity (CVSS 9.1), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
There is a possible out of bounds write due to a missing bounds check.Product: AndroidVersions: Android SoCAndroid ID: A-162980455. Rated critical severity (CVSS 9.1), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
There is a possible out of bounds read due to a missing bounds check.Product: AndroidVersions: Android SoCAndroid ID: A-162980705. Rated critical severity (CVSS 9.1), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
There is a possible out of bounds write due to a missing bounds check.Product: AndroidVersions: Android SoCAndroid ID: A-163008257. Rated critical severity (CVSS 9.1), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
An exploitable denial of service vulnerability exists in the ENIP Request Path Data Segment functionality of Allen-Bradley Flex IO 1794-AENT/B. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available and no vendor patch available.
An exploitable denial of service vulnerability exists in the ENIP Request Path Data Segment functionality of Allen-Bradley Flex IO 1794-AENT/B. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available and no vendor patch available.
An exploitable denial of service vulnerability exists in the ENIP Request Path Port Segment functionality of Allen-Bradley Flex IO 1794-AENT/B. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available and no vendor patch available.
An attacker who convinces a valid user to open a specially crafted project file to exploit could execute code under the privileges of the application due to an out-of-bounds read vulnerability on the. Rated high severity (CVSS 7.8), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
A denial of service vulnerability exists in the D3DKMTEscape handler functionality of AMD ATIKMDAG.SYS (e.g. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. This Out-of-bounds Read vulnerability could allow attackers to read data from memory outside the intended buffer boundaries.
A denial of service vulnerability exists in the D3DKMTCreateAllocation handler functionality of AMD ATIKMDAG.SYS (e.g. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. This Out-of-bounds Read vulnerability could allow attackers to read data from memory outside the intended buffer boundaries.
This vulnerability allows remote attackers to execute arbitrary code on affected installations of Foxit Reader 10.0.0.35798. Rated high severity (CVSS 7.8), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
This vulnerability allows remote attackers to execute arbitrary code on affected installations of Foxit PhantomPDF 10.0.0.35798. Rated high severity (CVSS 7.8), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
This vulnerability allows remote attackers to execute arbitrary code on affected installations of Foxit PhantomPDF 10.0.0.35798. Rated high severity (CVSS 7.8), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
This vulnerability allows remote attackers to disclose sensitive information on affected installations of Foxit PhantomPDF 10.0.0.35798. Rated low severity (CVSS 3.3), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
This vulnerability allows remote attackers to execute arbitrary code on affected installations of Microhard Bullet-LTE prior to v1.2.0-r1112. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
Integer Overflow or Wraparound vulnerability in the XML RPC library of OpenRobotics ros_comm communications packages allows unauthenticated network traffic to cause unexpected behavior. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity.
Taurus-AN00B versions earlier than 10.1.0.156(C00E155R7P2) have a buffer overflow vulnerability. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
Taurus-AN00B versions earlier than 10.1.0.156(C00E155R7P2) have a buffer overflow vulnerability. Rated medium severity (CVSS 6.5), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
HUAWEI P30 Pro versions earlier than 10.1.0.160(C00E160R2P8) and versions earlier than 10.1.0.160(C01E160R2P8) have a buffer overflow vulnerability. Rated high severity (CVSS 7.8), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
HUAWEI P30 Pro versions earlier than 10.1.0.160(C00E160R2P8) have an out-of-bounds read and write vulnerability. Rated medium severity (CVSS 5.5), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
HUAWEI P30 Pro versions earlier than 10.1.0.160(C00E160R2P8) have an out-of-bounds read and write vulnerability. Rated medium severity (CVSS 5.5), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
Taurus-AN00B versions earlier than 10.1.0.156(C00E155R7P2) have an out-of-bounds read and write vulnerability. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
Taurus-AL00A version 10.0.0.1(C00E1R1P1) has an out-of-bounds read vulnerability in XFRM module. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
A vulnerability in SonicOS allows a remote unauthenticated attacker to cause Denial of Service (DoS) on the firewall SSLVPN service by sending a malicious HTTP request that leads to memory addresses. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
A Heap Overflow vulnerability in the SonicOS allows a remote unauthenticated attacker to cause Denial of Service (DoS) on the firewall SSLVPN service and leads to SonicOS crash. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
A buffer overflow vulnerability in SonicOS allows a remote unauthenticated attacker to cause Denial of Service (DoS) on the firewall SSLVPN service and leads to firewall crash. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
A buffer overflow vulnerability in SonicOS allows an authenticated attacker to cause Denial of Service (DoS) in the SSL-VPN and virtual assist portal, which leads to a firewall crash. Rated medium severity (CVSS 6.5), this vulnerability is remotely exploitable, low attack complexity. No vendor patch available.
A buffer overflow vulnerability in SonicOS allows a remote attacker to cause Denial of Service (DoS) and potentially execute arbitrary code by sending a malicious request to the firewall. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
A vulnerability in SonicOS allows an authenticated attacker to cause out-of-bound invalid file reference leads to a firewall crash. Rated medium severity (CVSS 6.5), this vulnerability is remotely exploitable, low attack complexity. No vendor patch available.
A vulnerability in SonicOS allows a remote unauthenticated attacker to cause Denial of Service due to buffer overflow, which leads to a firewall crash. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
Certain NETGEAR devices are affected by a stack-based buffer overflow by an authenticated user. Rated medium severity (CVSS 6.8), this vulnerability is low attack complexity. No vendor patch available.
IBM Informix spatial 14.10 could allow a local user to execute commands as a privileged user due to an out of bounds write vulnerability. Rated high severity (CVSS 7.8), this vulnerability is low attack complexity. This Out-of-bounds Write vulnerability could allow attackers to write data beyond allocated buffer boundaries leading to code execution or crashes.
A vulnerability in the Cisco Discovery Protocol implementation for Cisco Video Surveillance 8000 Series IP Cameras could allow an unauthenticated, adjacent attacker to execute arbitrary code on an. Rated high severity (CVSS 8.8), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
Multiple buffer overflow vulnerabilities were found in the QUIC image decoding process of the SPICE remote display system, before spice-0.14.2-1. Rated medium severity (CVSS 6.6), this vulnerability is remotely exploitable, low attack complexity. This Buffer Copy without Size Check vulnerability could allow attackers to overflow a buffer to corrupt adjacent memory.
A stack overflow in WhatsApp for Android prior to v2.20.196.16, WhatsApp Business for Android prior to v2.20.196.12, WhatsApp for iOS prior to v2.20.90, WhatsApp Business for iOS prior to v2.20.90,. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
A buffer overflow in WhatsApp for Android prior to v2.20.130 and WhatsApp Business for Android prior to v2.20.46 could have allowed an out-of-bounds write when processing malformed local videos with. Rated high severity (CVSS 7.8), this vulnerability is low attack complexity. No vendor patch available.
The PPP implementation of MPD before 5.9 allows a remote attacker who can send specifically crafted PPP authentication message to cause the daemon to read beyond allocated memory buffer, which would. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available.
The L2TP implementation of MPD before 5.9 allows a remote attacker who can send specifically crafted L2TP control packet with AVP Q.931 Cause Code to execute arbitrary code or cause a denial of. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available.
A flaw was found in the HDLC_PPP module of the Linux kernel in versions before 5.9-rc7. Rated high severity (CVSS 7.2), this vulnerability is remotely exploitable, low attack complexity.
An issue was discovered in the box application on HiSilicon based IPTV/H.264/H.265 video encoders. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available and no vendor patch available.
The TCOS smart card software driver in OpenSC before 0.21.0-rc1 has a stack-based buffer overflow in tcos_decipher. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. This Out-of-bounds Write vulnerability could allow attackers to write data beyond allocated buffer boundaries leading to code execution or crashes.
The gemsafe GPK smart card software driver in OpenSC before 0.21.0-rc1 has a stack-based buffer overflow in sc_pkcs15emu_gemsafeGPK_init. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
The Oberthur smart card software driver in OpenSC before 0.21.0-rc1 has a heap-based buffer overflow in sc_oberthur_read_file. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. This Out-of-bounds Write vulnerability could allow attackers to write data beyond allocated buffer boundaries leading to code execution or crashes.
NVIDIA Virtual GPU Manager contains a vulnerability in the vGPU plugin and the host driver kernel module, in which the potential exists to write to a memory location that is outside the intended. Rated high severity (CVSS 7.1), this vulnerability is low attack complexity. No vendor patch available.
An issue was discovered in the client side of Zoho ManageEngine Desktop Central 10.0.0.SP-534. Rated high severity (CVSS 7.2), this vulnerability is remotely exploitable, low attack complexity. No vendor patch available.
NVIDIA Windows GPU Display Driver, all versions, contains a vulnerability in the DirectX11 user mode driver (nvwgf2um/x.dll), in which a specially crafted shader can cause an out of bounds access,. Rated high severity (CVSS 7.8), this vulnerability is low attack complexity. No vendor patch available.
The MSI AmbientLink MsIo64 driver 1.0.0.8 has a Buffer Overflow (0x80102040, 0x80102044, 0x80102050,and 0x80102054). Rated high severity (CVSS 7.8), this vulnerability is low attack complexity. Public exploit code available and no vendor patch available.
A remote buffer overflow vulnerability in the /cgi-bin/makeRequest.cgi endpoint of the WAVLINK WN530H4 M30H4.V5030.190403 allows an attacker to execute arbitrary machine instructions as root without. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
An issue was discovered in Foxit Reader and PhantomPDF before 10.1. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. This Out-of-bounds Write vulnerability could allow attackers to write data beyond allocated buffer boundaries leading to code execution or crashes.
An issue was discovered in Foxit Reader and PhantomPDF before 10.1. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. This Out-of-bounds Write vulnerability could allow attackers to write data beyond allocated buffer boundaries leading to code execution or crashes.
Artifex MuPDF before 1.18.0 has a heap based buffer over-write when parsing JBIG2 files allowing attackers to cause a denial of service. Rated medium severity (CVSS 5.5), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
When processing surfaces, the lifetime may outlive a persistent buffer leading to memory corruption and a potentially exploitable crash. Rated high severity (CVSS 8.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
Mozilla developers reported memory safety bugs present in Firefox 80. Rated high severity (CVSS 8.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
Mozilla developers reported memory safety bugs present in Firefox 80 and Firefox ESR 78.2. Rated high severity (CVSS 8.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
Mozilla developers reported memory safety bugs present in Firefox for Android 79. Rated high severity (CVSS 8.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
When processing a MAR update file, after the signature has been validated, an invalid name length could result in a heap overflow, leading to memory corruption and potentially arbitrary code. Rated high severity (CVSS 8.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
A vulnerability has been discovered in the ace.xmd parser that results from a lack of proper validation of user-supplied data, which can result in a write past the end of an allocated buffer. Rated high severity (CVSS 7.5), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
In PLC WinProladder Version 3.28 and prior, a stack-based buffer overflow vulnerability can be exploited when a valid user opens a specially crafted file, which may allow an attacker to remotely. Rated high severity (CVSS 7.8), this vulnerability is no authentication required, low attack complexity. No vendor patch available.
A flaw was found in dpdk in versions before 18.11.10 and before 19.11.5. Rated high severity (CVSS 8.8), this vulnerability is low attack complexity. This Buffer Copy without Size Check vulnerability could allow attackers to overflow a buffer to corrupt adjacent memory.
A flaw was found in dpdk in versions before 18.11.10 and before 19.11.5. Rated high severity (CVSS 7.1), this vulnerability is low attack complexity. This Out-of-bounds Read vulnerability could allow attackers to read data from memory outside the intended buffer boundaries.
A flaw was found in dpdk in versions before 18.11.10 and before 19.11.5. Rated high severity (CVSS 7.8). This Buffer Copy without Size Check vulnerability could allow attackers to overflow a buffer to corrupt adjacent memory.
url.cpp in libproxy through 0.4.15 is prone to a buffer overflow when PAC is enabled, as demonstrated by a large PAC file that is delivered without a Content-length header. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
A vulnerability in the Trend Micro Apex One ServerMigrationTool component could allow an attacker to trigger an out-of-bounds red information disclosure which would disclose sensitive information to. Rated medium severity (CVSS 4.3), this vulnerability is remotely exploitable, no authentication required, low attack complexity. No vendor patch available.
An out-of-bounds read information disclosure vulnerabilities in Trend Micro Apex One may allow a local attacker to disclose sensitive information to an unprivileged account on vulnerable. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
An out-of-bounds read information disclosure vulnerabilities in Trend Micro Apex One may allow a local attacker to disclose sensitive information to an unprivileged account on vulnerable. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
An out-of-bounds read information disclosure vulnerabilities in Trend Micro Apex One may allow a local attacker to disclose sensitive information to an unprivileged account on vulnerable. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
An out-of-bounds read information disclosure vulnerabilities in Trend Micro Apex One may allow a local attacker to disclose sensitive information to an unprivileged account on vulnerable. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
An out-of-bounds read information disclosure vulnerabilities in Trend Micro Apex One may allow a local attacker to disclose sensitive information to an unprivileged account on vulnerable. Rated medium severity (CVSS 5.5), this vulnerability is low attack complexity. No vendor patch available.
In TensorFlow Lite before versions 2.2.1 and 2.3.1, models using segment sum can trigger a write out bounds / segmentation fault if the segment ids are not sorted. Rated high severity (CVSS 8.1), this vulnerability is remotely exploitable, no authentication required. Public exploit code available.
In TensorFlow Lite before versions 2.2.1 and 2.3.1, models using segment sum can trigger a denial of service by causing an out of memory allocation in the implementation of segment sum. Rated medium severity (CVSS 4.0), this vulnerability is remotely exploitable, no authentication required. Public exploit code available.
In TensorFlow Lite before versions 2.2.1 and 2.3.1, models using segment sum can trigger writes outside of bounds of heap allocated buffers by inserting negative elements in the segment ids tensor. Rated high severity (CVSS 8.6), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available.
In TensorFlow Lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, saved models in the flatbuffer format use a double indexing scheme: a model has a set of subgraphs, each subgraph has a set. Rated medium severity (CVSS 4.8), this vulnerability is remotely exploitable, no authentication required. Public exploit code available.
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, if a TFLite saved model uses the same tensor as both input and output of an operator, then, depending on the operator, we can. Rated medium severity (CVSS 6.5), this vulnerability is remotely exploitable, no authentication required. Public exploit code available.
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, when determining the common dimension size of two tensors, TFLite uses a `DCHECK` which is no-op outside of debug compilation. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available.
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, to mimic Python's indexing with negative values, TFLite uses `ResolveAxis` to convert negative values to positive indices. Rated critical severity (CVSS 9.0), this vulnerability is remotely exploitable, no authentication required. Public exploit code available.
In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, the `data_splits` argument of `tf.raw_ops.StringNGrams` lacks validation. Rated critical severity (CVSS 9.8), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available.
In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, the `Shard` API in TensorFlow expects the last argument to be a function taking two `int64` (i.e., `long long`) arguments. Rated critical severity (CVSS 9.0), this vulnerability is remotely exploitable, no authentication required. Public exploit code available.
In Tensorflow before version 2.3.1, the `RaggedCountSparseOutput` implementation does not validate that the input arguments form a valid ragged tensor. Rated medium severity (CVSS 4.8), this vulnerability is remotely exploitable, no authentication required. Public exploit code available.