Insufficient verification of data authenticity may allow a local attacker to bypass Secure Debug Unlock (SDU), potentially resulting in unauthorized debug access.
Export limit exceeded: 426276 CVEs match your query. Please refine your search to export 10,000 CVEs or fewer.
Search vulnerabilities
Expected behavior violation in AMD Secure Encrypted Virtualization (SEV) firmware could allow a malicious hypervisor to bypass Reverse Map Table (RMP) protections, potentially resulting in a loss of Secure Encrypted Virtualization-Secure Nested Paging (SEV-SNP) guest memory integrity.
Improper Validation of Input in AMD SEV Firmware Allows Multiple RMP Entries, Compromising Guest Integrity
Improper validation of consistency within input in AMD SEV (Secure Encrypted Virtualization) firmware could allow an attacker with administrative privileges to create multiple RMP entries for the same guest physical address, potentially resulting in loss of guest integrity.
Insufficient Granularity of Access Control in AMD Secure Encrypted Virtualization (SEV) firmware could allow an attacker with administrative privileges attacker to write to a trusted region used by AMD Secure Encrypted Virtualization-Encrypted State (SEV-ES), potentially resulting in loss of integrity.
Incorrect usage of seeds in pseudo-random number generator in the Secure Encrypted Virtualization (SEV) firmware could allow a malicious hypervisor or migration agent to import a forged guest context potentially resulting in loss of confidentiality for derived guest keys.
Stored XSS in @waline/client via the unsanitized markdown preview of a comment's raw body on edit
Waline is a comment system with backend support. In versions 3.15.2 and prior, @waline/client renders the comment editor's live preview with marked and no sanitizer (utils/markdown.ts), and the preview is injected with v-html. When a comment is edited, the editor is seeded with the comment's raw, un-sanitized body (orig). Only administrators and the comment owner can edit. So an administrator who clicks Edit on an attacker's comment renders that attacker's raw HTML in their own session: an unauthenticated commenter achieves privilege-escalation stored XSS against a moderator with a single Edit click. At time of publication, there are no publicly available patches.
pam_usb: deny_remote bypassed by zone-indexed IPv6 addresses (fe80::1%eth0) in pusb_tmux_has_remote_clients
pam_usb provides hardware authentication for Linux using removable media. From version 0.8.0 to before version 0.9.3, the deny_remote feature in pam_usb uses /usr/bin/w -i to detect whether the authenticating user has an active remote tmux session. In pusb_tmux_has_remote_clients() (src/tmux.c), the IPv6 detection pattern matched the address group and then immediately expected whitespace. On Linux systems with multi-homed network interfaces, w -i appends a zone index (scope identifier) to link-local IPv6 addresses — e.g. fe80::1%eth0. The %eth0 suffix appeared between the address and the next whitespace field, causing the regex match to fail even for an otherwise-matching remote session. The function returned "no remote clients" and granted access. This issue has been patched in version 0.9.3.
pam_usb: getenv() in pusb_local_login bypasses XRDP deny via env-var unset
pam_usb provides hardware authentication for Linux using removable media. Prior to version 0.9.3, src/local.c reads XRDP_SESSION with plain getenv(3) at the head of pusb_local_login while every other env consumer in the same function uses secure_getenv(3). When pam_usb.so loads into a setuid PAM host, the calling user controls the environment passed in. Clearing XRDP_SESSION before invoking the host bypasses the explicit XRDP-deny rule. Other remote-detection signals (rmsvc, evdev, loginctl, utmp) may still catch the session, but the env-based detection alone is defeated. This issue has been patched in version 0.9.3.
pam_usb: Heap out-of-bounds read in pusb_get_tty_from_display_server cmdline scan
pam_usb provides hardware authentication for Linux using removable media. Prior to version 0.9.3, src/local.c (pusb_get_tty_from_display_server) allocates a 4096-byte cmdline buffer, reads up to 4096 bytes from /proc//cmdline, then replaces every NUL separator with a space. When read(2) returns exactly 4096 bytes for any process the function inspects, the buffer ends without a terminator and the subsequent strstr() calls scan past the allocation boundary. The function is invoked during normal PAM authentication when DISPLAY is set, so the read fires for every process at auth time. Consequences are bounded: strstr output is a Boolean substring match (limited info-disclosure surface) but a hostile heap layout can produce a SIGSEGV that terminates the host PAM process. This issue has been patched in version 0.9.3.
Omnispace Agora Project 26.10.0 Stored XSS via CSV Import Preview
Omnispace Agora Project through 26.10.0 contains a stored cross-site scripting vulnerability that allows attackers to inject malicious HTML markup by controlling imported directory data or an uploaded CSV file, as values are rendered into table cell text and value attributes after only Txt::utf8Encode() with no HTML escaping in VuePersonsImportExport.php. Attackers can craft a CSV cell containing an attribute-breakout payload that is rendered verbatim and executed in the browser of the administrator performing the import, leading to arbitrary script execution in the administrative context.
PostgreSQL type confusion after ALTER ... SET DATA TYPE executes arbitrary code
Type confusion in PostgreSQL allows an object creator to execute arbitrary code as the operating system user running the database, via ALTER ... SET DATA TYPE primarily. Other DDL attack routes exist, such as "ALTER TABLE ... ADD COLUMN; ROLLBACK; ALTER TABLE ... ADD COLUMN". The system had inadequate tracking of typed values, and values were not self-identifying. Stale values could appear in data dictionary entries, PL/pgSQL variables, and many other parts of the system. PostgreSQL Versions before PostgreSQL 18.7, 17.12, 16.16, 15.20, and 14.25 are affected.
In SetDisplayName of RegInfoContact.cpp, there is a possible denial of service due to a missing null check. This could lead to remote denial of service with no additional execution privileges needed. User interaction is not needed for exploitation.
Logic Error in Android Hardware Codec Firmware Enables Local Information Disclosure
In userdata_report of hwcodec_firmware, there is a possible Logic error due to a logic error in the code. This could lead to local information disclosure with execution privileges needed. User interaction is for exploitation.
Bahmni-Core: Path traversal may lead to arbitrary file write in the admin CSV import
OpenMRS module bahmnicore contains Core OpenMRS modules for Bahmni (including ERP & ELIS Atom Feed Clients). In version 0.93, there is insufficient validation of input paths during the admin CSV import process, allowing for potential arbitrary file write. This issue has been patched in version 0.93.1.
AIOHTTP: WebSocket max_msg_size bypass via integer overflow
AIOHTTP is an asynchronous HTTP client/server framework for asyncio and Python. Prior to 3.14.4, the C WebSocket reader adds a declared frame length to the size of an already buffered partial message using signed 64-bit arithmetic. A length near two to the power of 63 can overflow, allowing the max_msg_size check to pass. A remote peer can send a small non-final fragment followed by an oversized message and cause unbounded memory consumption. The Python WebSocket reader is unaffected. This issue is fixed in version 3.14.4.
llama.cpp b1886–b7445 Race Condition Use-After-Free via llama-android.cpp
llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Android JNI wrapper where bench_1model() and free_1context() lack synchronization, allowing Thread A to operate on freed memory while Thread B concurrently frees the llama_context. Attackers can exploit this by performing heap spray with attacker-controlled data containing a fake vtable to hijack the vtable pointer at offset +0x30, causing llama_batch_allocr::clear() to dereference arbitrary memory and achieve remote code execution.
CCC Attestation SIG attested-tls-poc through commit 731ca82 TOCTOU via Handshake-Only Attestation
CCC Attestation SIG attested-tls-poc through commit 731ca82 contains a time-of-check time-of-use vulnerability because Evidence is appraised only once during the TLS handshake, without post-handshake re-attestation. Attackers with local access to the attesting endpoint can alter workload code, data, configuration, or attached devices like GPUs while peers keep trusting the original Evidence.
gittuf: File protection rules bypassed for filenames with special characters
In gittuf before 0.18.0, verification of file protection rules uses the quoted, escaped pathnames that Git prints by default for files with special characters (for example non-ASCII characters) instead of the raw paths. Such paths do not match patterns like file:internal/*, so the files are treated as unprotected, and a user who is not authorized by a file protection rule can change a protected file whose name contains special characters and still pass verification.
gittuf: Propagated controller policies not loaded during verification
In gittuf 0.10.0 through 0.17.0, the verification workflow does not load policy updates that are propagated from an upstream (controller) repository and recorded in the reference state log after the first entry being verified. Only the propagated policy in effect at the start of verification is applied, so later controller changes to global rules, such as threshold requirements and force-push blocks, are ignored and changes that these rules should block pass verification.
gittuf: Invalid intermediate policies trusted when the policy apply workflow is bypassed
In gittuf 0.4.0 through 0.16.0, when a policy update appears within the range of reference state log entries being verified, verification checks only the root of trust signatures of the new policy and not the signatures of its other metadata (rule files). A user trusted to push to the repository who bypasses gittuf policy apply and writes the policy reference and reference state log directly can make an invalid policy apply during verification, weakening or strengthening the rules for later changes. The update is still recorded in the reference state log.