| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Incorrect caching of authentication between different users of the qSnapper dbus service before version 1.3.3 allowed any local attacker to use dbus functions after a privileged users has authenticated for them. |
| pnpm is a package manager. Prior to 10.34.4 and 11.7.0, a crafted patch entry could resolve outside the configured patches directory and cause pnpm patch-remove to delete an arbitrary reachable file. This vulnerability is fixed in 10.34.4 and 11.7.0. |
| pnpm is a package manager. Prior to 10.34.4 and 11.8.0, pnpm accepts package names from the env lockfile configDependencies section and uses those names directly when creating config dependency symlinks under node_modules/.pnpm-config. A malicious repository can commit a crafted pnpm-lock.yaml whose env-lockfile document contains a traversal-shaped config dependency name. During pnpm install, pnpm installs the config dependency and creates a symlink at a path derived from that name. This vulnerability is fixed in 10.34.4 and 11.8.0. |
| Untrusted Java Deserialization in Apache OpenNLP SvmDoccatModel
Versions Affected:
before 3.0.0-M4 (libsvm document categorization module; introduced in
OPENNLP-1808 and only present on the 3.x line)
Description:
SvmDoccatModel.deserialize(InputStream) reads an attacker-controlled
stream with java.io.ObjectInputStream and calls readObject() without an
ObjectInputFilter installed. ObjectInputStream materialises every class
referenced in the stream before the resulting object is cast to
SvmDoccatModel, so the cast that follows readObject() executes only
after the foreign object graph has already been deserialised in full.
If a Java deserialization gadget chain is available on the consumer's
classpath, a crafted payload supplied to
deserialize() executes arbitrary code in the JVM that loads it. Apache
OpenNLP itself does not ship a known gadget chain, so the realistic
risk is to downstream applications that embed the libsvm module
alongside vulnerable transitive dependencies. The method is public and
static, so any caller can pass an untrusted stream to it directly.
The practical impact is remote code execution against processes that
load SvmDoccatModel instances from untrusted or semi-trusted origins.
Mitigation:
3.x users should upgrade to 3.0.0-M4.
Users who cannot upgrade immediately should treat all serialized
SvmDoccatModel streams as untrusted input unless their provenance is
verified, and should avoid invoking SvmDoccatModel.deserialize() on
streams supplied by end users or fetched from third-party sources
without integrity checks. |
| vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, a frontend-legal multi-request speculative decoding workload can cause the rejection sampler to produce a recovered token equal to the model vocabulary size boundary value, which is then converted to negative one when the engine selects the next live token for a request and is written back into the drafter's input ids; that out-of-vocabulary value is later consumed by the model's embedding and attention path and crashes the engine worker with a GPU device-side assertion. The same triggering request sequence is reachable through the public gRPC Generate and Abort endpoints, so a remote client that can send generation requests can crash the shared engine worker, aborting concurrent requests and causing a service-wide denial of service for other clients of the deployment until the worker is restarted. This issue is fixed in version 0.24.0. |
| vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, the structured_outputs.regex API parameter passes a user-supplied regular expression string directly to the grammar compiler backends with no compilation timeout; in the xgrammar backend the string reaches the regex compiler with no guard, and in the outlines backend the validation step blocks structural issues such as lookarounds and backreferences but performs no complexity analysis, so a pattern with nested quantifiers passes all checks and causes exponential state-space expansion, allowing a single request containing an adversarial regex to hang an inference worker indefinitely and deny service. This issue is fixed in version 0.24.0. |
| Coolify is an open-source and self-hostable tool for managing servers, applications, and databases. Prior to 4.0.0-beta.474, PostgreSQL healthcheck command generation used attacker-controlled database settings (postgres_user and postgres_db) in shell-form commands, allowing an authenticated user to inject commands executed in the database container. This issue is fixed in version 4.0.0-beta.474. |
| Coolify is an open-source and self-hostable tool for managing servers, applications, and databases. Prior to 4.0.0-beta.471, the database import Livewire component (app/Livewire/Project/Database/Import.php) allows client-controlled container and server properties to reach shell commands without locking or validation, allowing an authenticated user to inject commands through a database import container name. This issue is fixed in version 4.0.0-beta.471. |
| In MLflow versions prior to 3.14.0, when running with authentication enabled, the trace API endpoints lack proper authorization validators. This allows any authenticated user to bypass experiment-level authorization controls on all trace operations, including reading, deleting, and modifying traces on experiments they do not have permission to access. The issue arises from the `_before_request` handler, which does not register authorization validators for trace endpoints, resulting in requests proceeding without validation. This vulnerability can expose sensitive data, destroy audit logs, and allow unauthorized modifications. |
| Coder allows organizations to provision remote development environments via Terraform. Prior to versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2, Coder's OIDC callback checked `email_verified` with a direct Go `bool` type assertion. When an IdP returned the claim as a non-boolean (for example the string `"false"`) or omitted it, the assertion failed open and the email was treated as verified. Combined with an unconditional email-based account fallback, this enabled account takeover. The fix in versions 2.29.7, 2.32.7, 2.33.8, and 2.34.2 coerces `email_verified` across bool, string and numeric types (fail-closed) and blocks the email fallback when the matched user already has a different linked IdP subject. As a workaround, ensure the IdP returns `email_verified` as a native JSON boolean. The email-fallback linking issue has no configuration workaround; upgrading is required. |
| Access of resource using incompatible type ('type confusion') in Microsoft Edge (Chromium-based) allows an unauthorized attacker to execute code over a network. |
| Pillow is a Python imaging library. Prior to 12.3.0, PIL/PcfFontFile.py _load_bitmaps() read glyph dimensions from the PCF METRICS section and passed them directly to Image.frombytes() without calling Image._decompression_bomb_check(), allowing crafted PCF font data to cause excessive memory allocation. This issue is fixed in version 12.3.0. |
| Memory safety bugs present in Firefox 152.0.3. Some of these bugs showed evidence of memory corruption and we presume that with enough effort some of these could have been exploited to run arbitrary code. This vulnerability was fixed in Firefox 152.0.4. |
| Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.9.1, an Insecure Direct Object Reference (IDOR) vulnerability in /api/v1/responses endpoint allows an authenticated attacker to execute any flow belonging to another user by specifying the victim's flow ID in the request. This vulnerability is fixed in 1.9.1. |
| NVIDIA Container Toolkit for Linux contains a vulnerability where an attacker could cause a time-of-check time-of-use race condition. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, and data tampering. |
| HTTP::Tiny versions before 0.095 for Perl forward credential headers to cross-origin redirect targets.
When the server returns a 3xx redirect, `_maybe_redirect` follows the `Location:` header and `_prepare_headers_and_cb` re-merges the caller's `headers` argument into the new request, without checking whether the redirect target shares an origin with the original URL. Caller-supplied `Authorization`, `Cookie` and `Proxy-Authorization` headers are therefore re-sent to whatever host the redirect names, across scheme, host or port boundaries, and including `https` to `http` downgrades that expose them in plaintext on the wire.
The HTTP::Tiny POD note that "Authorization headers will not be included in a redirected request" applied only to the URL-userinfo Basic-auth path, not to headers passed explicitly by the caller. |
| cgltf version 1.15 and prior contain an integer overflow vulnerability in the cgltf_validate() function when validating sparse accessors that allows attackers to trigger out-of-bounds reads by supplying crafted glTF/GLB input files with attacker-controlled size values. Attackers can exploit unchecked arithmetic operations in sparse accessor validation to cause heap buffer over-reads in cgltf_calc_index_bound(), resulting in denial of service crashes and potential memory disclosure. |
| When a user invokes curl using a schemeless URL combined with
`--proto-default` sftp (or scp), a disconnect occurs between the tool layer
and libcurl. The tool layer incorrectly infers the URL scheme, which
erroneously bypasses the initialization of critical SSH security options like
CURLOPT_SSH_HOST_PUBLIC_KEY_SHA256 and CURLOPT_SSH_KNOWNHOSTS. Conversely, the
libcurl runtime successfully honors CURLOPT_DEFAULT_PROTOCOL and establishes
the connection via SFTP/SCP as specified. Because the tool layer skipped the
security configuration, these SSH host verification options are silently
omitted, causing curl to connect to an unverified SSH remote host without
throwing an error. |
| Trail of Bits fickling versions up to and including 0.1.10 do not include the Python standard library modules _posixsubprocess, site, and atexit in the UNSAFE_IMPORTS denylist (fickle.py). Because these modules are absent from the denylist, fickling's check_safety() function returns LIKELY_SAFE with zero findings for pickle payloads that invoke dangerous functions including _posixsubprocess.fork_exec (C-level process spawner capable of executing arbitrary binaries), site.execsitecustomize (executes arbitrary site customization code), and atexit._run_exitfuncs (triggers all registered exit handler callbacks). The fickling.load() API chains check_safety() into pickle.loads() as an explicit security gate; a LIKELY_SAFE verdict causes the payload to be deserialized and executed. This shares the same root cause as CVE-2026-22607 (cProfile), CVE-2025-67748 (pty), and CVE-2025-67747 (marshal/types). OvertlyBadEvals does not flag these modules because they are standard library imports. UnsafeImports does not flag them because they are not in the denylist. The UnusedVariables heuristic is defeated by the SETITEMS opcode pattern. |
| In Trail of Bits fickling versions up to and including 0.1.11, the UnsafeImportsML analysis pass unconditionally calls AnalysisContext.shorten_code(node) on every import node it inspects, regardless of whether the import is flagged as unsafe. This call registers the shortened code representation in the shared AnalysisContext.reported_shortened_code set. When the MLAllowlist analysis pass subsequently runs, it calls the same shorten_code() method, receives already_reported=True for every import, and executes a continue statement that skips its allowlist check entirely. This renders MLAllowlist dead code for all imports — it never evaluates whether an import is in the ML allowlist or not. The MLAllowlist pass was designed to catch imports of modules outside the known-safe ML ecosystem (torch, numpy, transformers, etc.) that slip past the UnsafeImports denylist. With MLAllowlist inoperative, any standard library module not in the UNSAFE_IMPORTS denylist can be invoked via pickle deserialization while fickling's check_safety() returns LIKELY_SAFE. The fickling.load() API chains check_safety() into pickle.loads() as an explicit security gate, meaning a LIKELY_SAFE verdict causes the payload to be deserialized and executed. The root cause is shared mutable state between independently-correct analysis passes — UnsafeImportsML works as designed in isolation, MLAllowlist works as designed in isolation, but the shared reported_shortened_code set causes UnsafeImportsML to poison MLAllowlist's deduplication logic. |
