Information Disclosure

Format string vulnerability in Secure Access client for macOS prior to version 14.50 allows authenticated local attackers to leak sensitive data from process memory via crafted logging input, potentially exposing secrets through log files. The vulnerability requires local access and valid user privileges but no user interaction. Patch is available from the vendor.

Authorization bypass in Clerk JavaScript SDKs allows authenticated users to proceed past combined authorization checks they should fail. When developers use has() or auth.protect() with multiple authorization dimensions (e.g., role + reverification, permission + billing feature, or billing plan + permission), the predicate incorrectly returns true for users who satisfy only a subset of the required conditions. Sessions and authentication remain secure, but gated actions may execute for under-privileged users. Patches released across all affected SDK packages (Core 2 and Core 3) with no API changes. No public exploit code identified at time of analysis, but the vulnerability is straightforward to trigger in production code patterns explicitly outlined in the vendor advisory.

Out-of-bounds read in the GnuTLS DTLS handshake reassembly logic lets remote unauthenticated attackers trigger an integer underflow by sending malformed handshake fragments with zero length and a non-zero offset, enabling information disclosure or denial of service against Red Hat-shipped GnuTLS (RHEL 6 through 10, OpenShift Container Platform 4, and Red Hat Hardened Images). Reported by Red Hat with a vendor patch available and errata released; no public exploit identified at time of analysis, and EPSS is very low (0.04%, 11th percentile) with SSVC exploitation status 'none'. CVSS 9.1 reflects high confidentiality and availability impact over the network at low complexity without authentication.

GnuTLS with OCSP verification enabled incorrectly accepts revoked server certificates when presented with specially crafted multi-record OCSP responses during TLS handshakes, allowing attackers to bypass certificate revocation checks and establish connections to compromised servers. The vulnerability requires high attack complexity and specific OCSP configuration, affecting Red Hat Enterprise Linux 6-10, Red Hat Hardened Images, and OpenShift Container Platform 4. No public exploit code or active exploitation has been identified at the time of analysis.

Certificate name-constraints bypass in GnuTLS allows a remote attacker to craft a leaf certificate whose Subject Alternative Name differs only in letter casing from a constrained dNSName or rfc822Name, causing GnuTLS to accept a certificate that its excludedSubtrees/permittedSubtrees policy should reject. The flaw stems from case-sensitive comparison of name-constraint labels and enables unauthorized access, impersonation, or information disclosure across software (notably Red Hat Enterprise Linux 6-10 and OpenShift Container Platform 4) that relies on GnuTLS for X.509 validation. SSVC notes a public proof-of-concept exists, though EPSS exploitation probability is very low (0.03%, 9th percentile) and the issue is not in CISA KEV.

Weblate fails to revoke Django REST Framework API tokens when users change their passwords, allowing attackers with knowledge of a user's old credentials to continue accessing APIs indefinitely even after password changes. The vulnerability affects authenticated users on Weblate versions prior to 5.17.1 and requires high attack complexity due to the need for legitimate account access, but carries meaningful risk in multi-user translation environments where password compromise may go undetected.

Keycloak's Account REST API remains partially accessible even when explicitly disabled via the `--features-disabled=account,account-api` flag, allowing authenticated users to read and modify account data through five unprotected endpoints under `/account/v1alpha1/` that lack the required `checkAccountApiEnabled()` access control gate present in four sibling endpoints within the same service class.

Authenticated users with minimal namespace-scoped privileges can obtain administrative credentials for arbitrary OpenShift clusters provisioned through the MCE hub via the assisted-service REST API. The vulnerability exists in AUTH_TYPE=local mode (the only mode available in on-premises deployments), where the local authenticator grants full administrative access to any request bearing a valid JWT with no per-endpoint restrictions. A valid JWT is embedded as plaintext in the InfraEnvStatus.ISODownloadURL, readable by any user with get rights on an InfraEnv object, enabling extraction of kubeadmin passwords and kubeconfigs for all spoke clusters.

Remote authenticated attackers can flood PDKS (Personnel and Document Tracking System) through uncontrolled interaction frequency, achieving high integrity and availability impacts without confidentiality breach. This workforce management software by MeWare Software Development Inc. is vulnerable to denial-of-service conditions and potential data integrity compromise through rate-limiting bypass. Affects versions from V16.20200313 through VMYR_3.5.2025117. TR-CERT advisory available, EPSS data not provided, no CISA KEV listing identified.

Code injection in Qt SVG module allows attackers to execute arbitrary QML/JavaScript when applications load malicious SVG files through Qt Quick's VectorImage component. Exploitation requires local file access and user interaction (opening crafted SVG). While QML execution is more restricted than native code, attackers can still trigger denial of service, exfiltrate application data, or manipulate UI logic depending on the victim application's privilege context. No active exploitation confirmed (not in CISA KEV), but patch available from Qt Project reduces urgency for immediate emergency response.

Authenticated remote attackers can access sensitive personal information in MeWare PDKS versions 16.20200313 through before VMYR_3.5.2025117 due to improper access controls. The vulnerability allows disclosure of private data without authorization, affecting confidentiality. CVSS score of 6.5 reflects moderate severity with network accessibility and low attack complexity, though authentication is required. No active exploitation or public proof-of-concept has been confirmed at time of analysis.

Dancer::Session::Abstract through version 1.3522 generates cryptographically weak session identifiers by combining predictable inputs (file path, process ID, epoch time) with an insufficiently-seeded Perl rand() function, allowing remote attackers to predict valid session IDs and hijack user sessions without authentication. The vulnerability affects Perl-based web applications using Dancer framework's default session handling; active exploitation is not confirmed but the attack requires only guessing a session ID, making it practically exploitable.

Uninitialized variable use in Linux kernel CIFS replay logic allows local authenticated attackers to potentially access sensitive kernel memory, corrupt data, or trigger denial of service. The vulnerability exists in CIFS request replay code paths where certain local variables lack proper reinitialization after replay labels, potentially causing undefined behavior during SMB session recovery operations. Patches available for kernel versions 6.6.128, 6.12.75, 6.18.16, 6.19.6, and 7.0. EPSS score of 0.02% indicates minimal observed exploitation activity, consistent with the local access requirement and specialized triggering conditions.

Local file disclosure in IntelliJ IDEA's built-in web server allows remote attackers to read arbitrary local files via network requests requiring user interaction. JetBrains IDEA versions before 2024.3.7.1, 2025.1.7.1, 2025.2.6.2, 2025.3.4.1, and 2026.1.1 are affected. The vulnerability achieves scope change (CVSS S:C) enabling cross-context information theft with high confidentiality impact but no integrity or availability damage. No active exploitation (KEV-absent) or public POC identified at time of analysis, though vendor disclosure suggests controlled remediation timeline.

In the Linux kernel, the following vulnerability has been resolved: rtnetlink: add missing netlink_ns_capable() check for peer netns rtnl_newlink() lacks a CAP_NET_ADMIN capability check on the peer network namespace when creating paired devices (veth, vxcan, netkit). This allows an unprivileged user with a user namespace to create interfaces in arbitrary network namespaces, including init_net. Add a netlink_ns_capable() check for CAP_NET_ADMIN in the peer namespace before allowing device creation to proceed.

Double-free vulnerability in Linux kernel Xen privcmd driver allows local authenticated attackers to corrupt memory and potentially execute arbitrary code or cause denial of service. When userspace performs partial munmap() on privcmd mappings, VMA splitting creates duplicate pointers to the same memory pages array, leading to kvfree() being called twice on the same allocation during VMA cleanup. Xen Security Advisory XSA-487 confirms this issue affects virtualization hosts running Xen paravirtualized domains. No public exploit identified at time of analysis, with EPSS score of 0.03% indicating low predicted exploitation probability. Vendor-released patches available for stable kernel versions 5.10.254, 5.15.204, 6.1.170, 6.6.137, 6.12.85, 6.18.26, and 7.0.3.

Apache Airflow's SmtpHook performs STARTTLS upgrades without SSL certificate validation, allowing man-in-the-middle attackers to intercept SMTP credentials. Remote unauthenticated attackers positioned between an Airflow worker and SMTP server can present a self-signed certificate, complete the TLS handshake, and capture login credentials sent after the upgrade. The vulnerability affects apache-airflow-providers-smtp versions 2.0.0 through 2.x and is patched in version 3.0.0 or later. No public exploit code identified at time of analysis, but EPSS score of 0.01% indicates low real-world exploitation probability despite confidentiality impact.

Out-of-bounds read in ASR Kestrel's nr_fw power control module enables authenticated remote attackers to trigger buffer overflow conditions, potentially disclosing sensitive information and compromising system integrity with low-to-moderate impact across security boundaries. Exploitable over the network with low complexity by attackers holding low-privilege credentials. EPSS data unavailable; not currently listed in CISA KEV. Vendor advisory confirms patch released February 10, 2026.

Out-of-bounds read in Exim before 4.99.2 when UTF-8 operators are enabled allows remote unauthenticated attackers to leak sensitive information through error messages by sending email with malformed UTF-8 trailing characters in headers. The vulnerability has high attack complexity due to the requirement for UTF-8 operator enablement and specific malformed input crafting, but requires no user interaction and operates over the network on default deployments.

The `BetaLocalFilesystemMemoryTool` in the Anthropic TypeScript SDK created memory files and directories using the Node.js default modes (`0o666` for files, `0o777` for directories), leaving them world-readable on systems with a standard umask and world-writable in environments with a permissive umask such as many Docker base images. A local attacker on a shared host could read persisted agent state, and in containerized deployments could modify memory files to influence subsequent model behavior. Users on the affected versions are advised to update to the latest version. Claude SDK thanks `lucasfutures` for the report.

Plack::Middleware::XSendfile through version 1.0053 allows remote unauthenticated attackers to read arbitrary files from nginx-proxied servers by injecting malicious X-Sendfile-Type and X-Accel-Mapping headers. When the middleware's sendfile type is not explicitly configured, clients can force nginx's X-Accel-Redirect mode and manipulate path mappings to access sensitive files outside intended directories. The middleware has been deprecated as of version 1.0053 and will be removed in future Plack releases. EPSS score of 0.01% suggests low current exploitation activity despite the high CVSS 9.1 rating. No public exploit code identified at time of analysis, though the attack technique mirrors the documented CVE-2025-61780 vulnerability in Rack::Sendfile.

SAML response redirection in Admidio 5.0.8 and earlier allows attackers to steal signed authentication assertions containing user credentials and profile data by exploiting missing validation of the AssertionConsumerServiceURL in SAML AuthnRequests. The Identity Provider (IdP) accepts attacker-controlled destination URLs from SAML requests without verifying them against registered Service Provider endpoints, enabling assertion theft and account impersonation across federated applications. No public exploit code identified at time of analysis, though proof-of-concept demonstration is included in the GitHub security advisory GHSA-p9w9-87c8-m235. EPSS data not available for this CVE.

Admidio members_assignment_data.php endpoint leaks hidden profile field values through a blind search oracle attack. Role leaders with ROLE_LEADER_MEMBERS_ASSIGN permissions can infer exact values of hidden PII fields (birthdays, street addresses, cities, postal codes, countries) by observing which users appear in search results, despite these fields being suppressed from JSON output. The vulnerability affects Admidio versions up to 5.0.8 and is fixed in 5.0.9.

Prototype pollution in n8n's XML node allows authenticated workflow editors to achieve remote code execution through global prototype manipulation. The vulnerability affects n8n workflow automation platform versions prior to 1.123.32, 2.17.4, and 2.18.1, enabling attackers with workflow creation privileges to inject malicious properties into JavaScript object prototypes that can be exploited by other nodes to execute arbitrary code. Vendor-released patches are available for all affected version branches. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis, though the CVSS 10.0 score reflects the critical scope change and complete system compromise potential.

CKAN fails to validate SMTP server certificates, allowing attackers to spoof the configured mail server with any certificate including self-signed ones and intercept SMTP credentials and email content via man-in-the-middle attack. Versions below 2.10.10 and 2.11.0 through 2.11.4 are affected. Vendor-released patches are available in CKAN 2.10.10 and 2.11.5.

GNU wget2 incorrectly validates TLS server certificates, accepting certificates with improper Key Usage (KU) or Extended Key Usage (EKU) attributes. This allows attackers who have compromised certificates issued for non-server purposes to impersonate legitimate servers in TLS connections, enabling man-in-the-middle attacks that leak sensitive information such as authentication credentials or request/response data.

Insufficient option length validation in the IPv6 Router Advertisement parser in FreeRTOS-Plus-TCP before V4.2.6 and V4.4.1 allows an adjacent network actor to cause a denial of service (device crash) by sending a crafted Router Advertisement with a truncated PREFIX_INFORMATION option that is smaller than the expected structure size. To mitigate this issue, users should upgrade to the fixed version when available.

Dell/Alienware Purchased Apps, versions prior to 1.1.31.0, contain an Improper Link Resolution Before File Access ('Link Following') vulnerability. A low privileged attacker with local access could potentially exploit this vulnerability, leading to Arbitrary File Write

AgentFlow's local web API accepts non-JSON content types on POST /api/runs and POST /api/runs/validate endpoints without enforcing application/json validation, allowing attackers to bypass trust-boundary enforcement on sensitive operations. Attackers can exploit this content-type validation weakness through browser-driven or local cross-origin requests to abuse the localhost API and enable attack chains against the local control plane.

A security flaw has been discovered in NousResearch hermes-agent 0.8.0. This affects the function _check_sensitive_path of the file tools/file_tools.py. The manipulation results in symlink following. Attacking locally is a requirement. The exploit has been released to the public and may be used for attacks. Upgrading to version 0.9.0 is able to mitigate this issue. The patch is identified as 311dac197145e19e07df68feba2cd55d896a3cd1. Upgrading the affected component is recommended.

An authorization flaw in the user management command could allow an authenticated user to make limited changes to authentication-related data associated with another user account. This could affect how authentication is performed for the impacted account.

Computing the MD5 checksum of a malformed BSON object under specific conditions may cause loss of availability in MongoDB server. This issue affects all MongoDB Server v8.2 versions, all MongoDB Server v8.1 versions, MongoDB Server v8.0 versions prior to 8.0.21, MongoDB Server v7.0 versions prior to 7.0.32

Netskope was notified about a potential gap in the Endpoint DLP Module for Netskope Client on Windows systems. The successful exploitation of the gap can potentially allow an unprivileged user to trigger an out-of-bounds read within a driver, leading to a Blue-Screen-of-Death (BSOD). Successful exploitation would require the Endpoint DLP module to be enabled in the client configuration. A successful exploit can potentially result in a denial-of-service for the local machine.

A vulnerability in the access control mechanism of SonicOS may allow certain management interface functions to be accessible under specific conditions.

Improper link resolution before file access ('link following') vulnerability in TUBITAK BILGEM Software Technologies Research Institute Pardus About allows Symlink Attack. This issue affects Pardus About: before v1.2.1.

Jenkins Matrix Authorization Strategy Plugin 2.0-beta-1 through 3.2.9 (both inclusive) invokes parameterless constructors of classes specified in configuration when deserializing inheritance strategies, without restricting the classes that can be instantiated, allowing attackers with Item/Configure permission to instantiate arbitrary types, which may lead to information disclosure or other impacts depending on the classes available on the classpath.

Ollama for Windows does not perform integrity or authenticity verification of downloaded update executables. Unlike other platforms, the Windows implementation of the update verification routine unconditionally returns success so no digital signature or trust validation is performed before staging or executing update payloads, enabling attacker‑supplied executables to be accepted and later executed by the application. Critically, Ollama for Windows performs silent automatic updates, so the malicious payload may be installed automatically without user awareness. Maintainers of this project were notified early about this vulnerability, but didn't respond with the details of vulnerability or vulnerable version range. Versions from 0.12.10 to 0.17.5 were tested and confirmed as vulnerable, other versions were not tested but might also be vulnerable.

Exposure of Sensitive System Information to an Unauthorized Control Sphere vulnerability in WPDeveloper BetterDocs betterdocs allows Retrieve Embedded Sensitive Data.This issue affects BetterDocs: from n/a through <= 4.3.10.

This vulnerability exists in e-Sushrut due to disclosure of sensitive information and hardcoded AES encryption keys in client-side JavaScript. An unauthenticated remote attacker could exploit this vulnerability by accessing the client-side code to extract sensitive information and cryptographic keys. Successful exploitation of this vulnerability could lead to exposure of sensitive data and compromise of cryptographic protections on the targeted system.

This vulnerability exists in e-Sushrut due to the use of reversible Base64 encoding for protecting sensitive data. An authenticated attacker could exploit this vulnerability by decoding and manipulating Base64-encoded parameters in the request URL to gain unauthorized access to sensitive information on the targeted system.

Insufficient verification of data authenticity in PackageManagerService prior to SMR Mar-2026 Release 1 allows local attackers to modify the installation restriction of specific application.

Dell iDRAC10, versions 1.20.70.50 and 1.30.05.10, contains an Insufficiently Protected Credentials vulnerability. A race condition vulnerability exists that could allow an authenticated low‑privileged attacker to gain elevated access.

Starman versions before 0.4018 for Perl allows HTTP Request Smuggling via Improper Header Precedence. Starman incorrectly prioritizes "Content-Length" over "Transfer-Encoding: chunked" when both headers are present in an HTTP request. Per RFC 7230 3.3.3, Transfer-Encoding must take precedence. An attacker could exploit this to smuggle malicious HTTP requests via a front-end reverse proxy.

Insufficient validation of untrusted input in Feedback in Google Chrome prior to 147.0.7727.138 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)

Race in MHTML in Google Chrome prior to 147.0.7727.138 allowed an attacker who convinced a user to install a malicious extension to leak cross-origin data via a crafted Chrome Extension. (Chromium security severity: High)

Out of bounds read and write in Angle in Google Chrome prior to 147.0.7727.138 allowed a remote attacker to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)

### Summary Fiber cache middleware's default key generator uses only `c.Path()` and does not include the query string. As a result, requests like `/?id=1` and `/?id=2` can map to the same cache key and share the same cached response. This can cause response mix-up (cache poisoning-like behavior) for endpoints where response content depends on query parameters. ### Details Default configuration in cache middleware: - `KeyGenerator: func(c fiber.Ctx) string { return utils.CopyString(c.Path()) }` References: - https://github.com/gofiber/fiber/blob/main/middleware/cache/config.go#L90-L92 - https://github.com/gofiber/fiber/blob/main/middleware/cache/cache_test.go#L599-L621 The existing test demonstrates that when handler output depends on query parameter `id`, a second request with a different query still returns the first cached response (cache hit), confirming query is not part of the default cache key. ### PoC Minimal PoC: ```go package main import ( "log" "github.com/gofiber/fiber/v3" "github.com/gofiber/fiber/v3/middleware/cache" ) func main() { app := fiber.New() app.Use(cache.New()) // default config app.Get("/", func(c fiber.Ctx) error { return c.SendString(c.Query("id", "1")) }) log.Fatal(app.Listen(":3000")) } ``` Reproduction: 1. `GET /?id=1` - Cache miss - Response body: `1` 2. `GET /?id=2` - Cache hit - Response body: `1` (expected `2`) Local verification command used: ```bash go test ./middleware/cache -run Test_Cache_WithNoCacheRequestDirective -count=1 ``` Observed result: test passes, confirming this is current behavior. ### Impact - Responses that should vary by query parameters can be mixed between requests. - In real deployments, this may leak or corrupt user/tenant-specific content if query parameters influence context or data selection. - This is deployment-dependent but security-relevant, and not safe-by-default for query-variant responses. ### Suggested remediation - Change default cache key generation to include path + normalized query string (or canonicalized original URL). - Keep ability for custom key generators. - Add explicit documentation warning that path-only keying is unsafe for query-dependent responses.

A security vulnerability has been detected in Xuxueli xxl-job up to 3.3.2. The impacted element is an unknown function of the file xxl-job-admin/src/main/java/com/xxl/job/admin/scheduler/openapi/OpenApiController.java of the component OpenAPI Endpoint. Such manipulation of the argument default_token leads to use of hard-coded cryptographic key . It is possible to launch the attack remotely. A high complexity level is associated with this attack. The exploitability is regarded as difficult. The exploit has been disclosed publicly and may be used.

A security flaw has been discovered in Xuxueli xxl-job up to 3.3.2. Impacted is the function logDetailCat of the file xxl-job-admin/src/main/java/com/xxl/job/admin/controller/biz/JobLogController.java of the component Execution Log Handler. The manipulation of the argument logId results in improper control of resource identifiers. The attack may be performed from remote. This attack is characterized by high complexity. The exploitability is considered difficult. The exploit has been released to the public and may be used for attacks. Upgrading to version 3.4.0 is recommended to address this issue. The patch is identified as d24e4ccd6073cc75305e1d3b9c29bc8db7437e7a. It is suggested to upgrade the affected component.

Improper Verification of Cryptographic Signature (CWE-347) in Elastic Package Registry could allow an attacker positioned to intercept network traffic, or to otherwise influence the contents served to a self-hosted registry, to substitute a tampered package without the integrity check failing closed.

OpenClaw versions before 2026.4.8 fail to enforce integrity verification on downloaded plugin archives. Attackers can install malicious or tampered plugin packages without detection, compromising the local assistant environment.

OpenClaw before 2026.4.8 treats shared reply MEDIA paths as trusted, allowing crafted references to trigger cross-channel local file exfiltration. Attackers can exploit this by crafting malicious shared reply MEDIA references to cause another channel to read local file paths as trusted generated media.

OpenClaw before 2026.4.8 fails to remove git plumbing environment variables from the execution environment before host exec operations. Attackers can exploit this by setting GIT_DIR and related variables to redirect git operations and compromise repository integrity.

OpenClaw before version 2026.4.2 leaks shared-secret length information through timing side-channel attacks in cryptographic comparison operations. The vulnerability stems from early length-mismatch checks in shared-secret comparison routines that violate constant-time security requirements, allowing remote attackers to measure timing differences and infer secret lengths without authentication. This weakens the cryptographic guarantees of the library's shared-secret handling.

OpenClaw before version 2026.3.31 accepts arbitrary tailnet peers as DNS authorities due to improper validation in its wide-area discovery mechanism, enabling attackers positioned within the same tailnet with CA-trusted endpoint access to manipulate DNS resolution and exfiltrate operator credentials. The vulnerability requires adjacent network access, high attack complexity, and user interaction, but results in high confidentiality impact through credential theft. No active exploitation has been publicly confirmed at the time of analysis.

OpenClaw package manager allows supply chain attacks through incomplete environment variable sanitization before version 2026.3.22. Attackers can hijack approved package installation or execution requests by injecting environment variables that redirect package resolution to malicious infrastructure, enabling trojanized code execution with high impact to confidentiality, integrity, and availability. This requires local access and user interaction to trigger package manager operations, limiting remote exploitation but creating significant insider threat and social engineering risk vectors.

Environment variable injection in OpenClaw's CLI backend runner enables local attackers to achieve arbitrary code execution or exfiltrate sensitive data by manipulating workspace configuration files. Attackers with the ability to supply malicious workspace configs can inject environment variables into backend processes during spawning, exploiting CWE-15 (external control of system or configuration setting). Vendor patch available via GitHub commit c2fb7f1. CVSS 8.5 reflects high impact across confidentiality, integrity, and availability, though exploitation requires local access and user interaction to load the malicious workspace config. No evidence of active exploitation (not in CISA KEV) or public proof-of-concept at time of analysis.

Execution approval bypass in OpenClaw before 2026.3.28 allows local authenticated users with standard privileges to establish overly broad executable allowlist entries through wrapper carrier exploitation. Attackers leverage positional routing in dispatch wrappers to trust carrier executables instead of their invoked targets, escalating from limited execution approval to arbitrary code execution with high confidentiality and integrity impact. Vendor-released patch version 2026.3.28 addresses the flaw (GHSA-p4x4-2r7f-wjxg). No evidence of active exploitation or public POC identified at time of analysis.

OpenClaw before version 2026.3.31 fails to block plugin installation when security scans detect threats, allowing authenticated users to install malicious plugins by ignoring visible scan warnings. The vulnerability requires user interaction (UI:P) and authenticated access (PR:L), but enables installation of untrusted code with moderate integrity impact when exploited.

Server-side request forgery in NVIDIA NemoClaw's validateEndpointUrl() function allows local attackers with user interaction to supply crafted endpoint URLs targeting the 0.0.0.0/8 address range via blueprint configuration files or CLI flags, leading to information disclosure. The vulnerability affects all versions of NemoClaw and requires local access with user interaction to trigger, limiting exposure to systems where untrusted users can modify configuration or invoke CLI commands.

Remote unauthenticated attackers can exfiltrate sensitive host environment variables from NVIDIA NeMoClaw by injecting malicious prompts that bypass sandbox access controls. The vulnerability affects the sandbox initialization component and enables information disclosure without requiring any authentication or user interaction (CVSS 8.6, AV:N/AC:L/PR:N/UI:N). Cross-scope impact (S:C) indicates the attack breaks out of the intended sandbox boundary to access host-level secrets. EPSS and KEV status not available; this appears to be a recently disclosed AI/LLM agent security issue.

NVIDIA Flare SDK is vulnerable to path traversal via improper input validation, allowing authenticated remote attackers to disclose sensitive information. The vulnerability affects all versions of the SDK and requires valid user credentials to exploit, making it a moderate-risk issue for organizations using Flare in multi-user environments. No public exploit code or active exploitation has been identified.

Authentication bypass in NVIDIA NVFlare Dashboard allows remote unauthenticated attackers to escalate privileges through user-controlled key manipulation in the authentication system. The vulnerability affects the NVIDIA Flare SDK and enables complete system compromise including arbitrary code execution, data tampering, information disclosure, and denial of service. With a CVSS score of 9.8 (critical severity) and maximum exploitability metrics (AV:N/AC:L/PR:N/UI:N), this represents a severe security flaw requiring immediate remediation, though no active exploitation (KEV) or public exploit code has been identified at time of analysis.

A vulnerability affecting the detailed versions of Cryptobox allows a legitimate user to prevent another to login by triggering an account lockout via sending a specially crafted request.

HTTP request smuggling in Apache Pony Mail (Lua implementation) enables remote unauthenticated attackers to achieve complete admin account takeover with critical impact across confidentiality, integrity, and availability. This affects all versions of the retired Lua codebase - Apache has abandoned support with no patch planned, recommending migration to alternative solutions. CVSS 9.8 critical severity reflects trivial network-based exploitation requiring no authentication or user interaction.

Spring gRPC versions 1.0.0 through 1.0.2 leak sensitive authentication failure details in gRPC status descriptions to unauthenticated remote callers, enabling reconnaissance for follow-up attacks. The vulnerability exposes raw server-side AuthenticationException messages without sanitization, providing attackers with information about authentication mechanisms and potential weaknesses. This low-severity information disclosure (CVSS 3.7) requires high attack complexity but affects default configurations.

Unauthenticated remote attackers can access sensitive documents containing personally identifiable information (PII) in Aranda Service Desk versions prior to 8.3.12 by exploiting predictable log file names in the Aranda File Server (AFS) component. Attackers retrieve daily activity logs from a publicly accessible directory to obtain direct virtual paths of uploaded files, then bypass access controls to download the documents. CISA SSVC framework confirms proof-of-concept code exists and the vulnerability is fully automatable, significantly lowering the barrier to exploitation despite no confirmed active exploitation at time of analysis.

GNU nano creates the ~/.local directory with world-writable permissions (0777) on first use of XDG data storage features when the directory does not exist, allowing local attackers in systems with relaxed umasks (such as containers, CI/CD runners, or environments with umask 000) to write attacker-controlled files into the victim's XDG directory hierarchy via a race condition. The vulnerability affects nano versions before 9.0 and carries a CVSS score of 2.1 with CISA SSVC assessment indicating partial technical impact but no known public exploitation.

Information disclosure in Mozilla Firefox, Firefox ESR 140, and Firefox ESR 115 allows remote unauthenticated attackers to extract sensitive data via incorrect boundary conditions in the Audio/Video component. The vulnerability permits network-based exploitation with low complexity and no user interaction (CVSS:3.1/AV:N/AC:L/PR:N/UI:N), enabling unauthorized access to high-confidentiality information. Mozilla released patches in Firefox 150.0.1, Firefox ESR 140.10.1, and Firefox ESR 115.35.1 (confirmed by vendor advisories MFSA2026-35/36/37). SSVC indicates automatable exploitation with partial technical impact, though no public exploit or active exploitation is identified at time of analysis.

OpenShift Container Platform build system allows authenticated users with the edit ClusterRole to inject arbitrary environment variables into docker-build containers via the buildconfigs/instantiate API, enabling information disclosure attacks such as build traffic interception through LD_PRELOAD or http_proxy manipulation. This represents an incomplete remediation of a prior vulnerability, affecting confidentiality of sensitive build data with CVSS 4.3 (network-accessible, low complexity, authenticated). No public exploit code or active exploitation has been confirmed at the time of analysis.

Remote command execution in mpGabinet 23.12.19 and below allows authenticated database administrators or unauthenticated attackers (via chained exploitation of CVE-2026-40550 and CVE-2026-40551) to achieve system command execution by manipulating attachment storage paths in the database to reference attacker-controlled resources that execute when users open the files. The vulnerability requires direct database access and user interaction to trigger execution, but becomes unauthenticated when chained with companion CVE vulnerabilities that grant database and application access.

Spring AI versions 1.0.0-1.0.5 and 1.1.0-1.1.4 expose ONNX machine learning models to unauthorized disclosure when the application runs in shared hosting environments, allowing local users with limited system access to read sensitive model files and potentially reverse-engineer proprietary ML logic. The vulnerability stems from insecure temporary file handling (CWE-377) that fails to restrict file permissions on extracted model artifacts. Authentication requirements are minimal-only local system access is needed-making this a significant risk in multi-tenant cloud platforms and shared servers.

Penetration Testing engineers at Amazon have discovered a flaw where the camera system fails to properly handle data supplied in certain requests, causing a service disruption. Rated medium severity (CVSS 5.3), this vulnerability is remotely exploitable, low attack complexity.

Out-of-bounds read in Artifex MuPDF up to version 1.28.0 within the CFF Index Handler's fz_subset_cff_for_gids function allows local attackers with low privileges to disclose sensitive information from application memory. The vulnerability requires local access and low privilege level but can be triggered without user interaction; publicly available exploit code exists and the vulnerability remains unpatched as of the last vendor response.

Command injection via ipmitool in OpenStack Ironic through 25.0.0 allows authenticated operators with high privileges to trigger ipmitool execution when a console interface is configured in a non-default deployment, leading to high impact on confidentiality, integrity, and availability of the bare-metal provisioning node. No public exploit identified at time of analysis, EPSS is very low (0.07%, 20th percentile), and SSVC indicates no observed exploitation, consistent with a niche operator-level flaw rather than mass-scanning risk.

Milesight AIOT cameras ship with hardcoded SSL private keys enabling remote man-in-the-middle attacks and credential interception. Remote unauthenticated attackers can decrypt TLS traffic, impersonate camera services, and potentially gain administrative access to affected devices. CISA ICS-CERT published advisory ICSA-26-113-03 for this industrial/IoT vulnerability affecting network-connected surveillance infrastructure.

Out-of-bounds read in Apache Thrift C++ JSON deserialization allows remote attackers to leak sensitive information and trigger denial of service via malformed JSON payloads. Affects Apache Thrift versions prior to 0.23.0. The vulnerability has low exploitation probability (EPSS 0.02%) and is not currently listed in CISA KEV, suggesting limited real-world weaponization despite network-accessible attack vector.

Out-of-bounds read vulnerability in Apache Thrift Swift implementation allows remote unauthenticated attackers to trigger denial of service and disclose limited memory contents via malformed skip() operations during protocol deserialization. Affects all versions prior to 0.23.0, with publicly disclosed exploit details on oss-security mailing list. EPSS exploitation probability remains low (5th percentile) despite network-accessible attack vector, suggesting limited real-world targeting to date. Vendor patch released in version 0.23.0 addresses all six concurrently disclosed Thrift vulnerabilities (CVE-2026-41602 through CVE-2026-41607).

Environment variable injection in OpenClaw (pre-2026.3.31) allows authenticated remote attackers to compromise host execution integrity by injecting malicious variables that override package managers, Docker registries, compiler paths, and TLS configurations during host exec operations. The vulnerability exhibits high confidentiality impact (CVSS:4.0 VC:H) with network attack vector and low complexity (AV:N/AC:L), requiring only low-privilege authentication (PR:L). VulnCheck disclosure indicates this affects Docker-related operations, with fixes available via GitHub commit eb8de67 and tracked under GHSA-cg7q-fg22-4g98. EPSS and KEV data not available at time of analysis.

Environment variable disclosure in OpenClaw jq safe-bin policy allows authenticated remote attackers to extract sensitive credentials and configuration data. The vulnerability stems from incomplete filter blocking in jq program execution - specifically, the $ENV filter can bypass safe-bin restrictions to read process environment variables. Versions prior to 2026.3.28 are affected. No CISA KEV listing or public POC identified at time of analysis, but disclosure by VulnCheck indicates vendor-confirmed issue with available patch.

OpenClaw before version 2026.3.31 allows authenticated attackers to read arbitrary host files through improper validation in the appendLocalMediaParentRoots function, enabling exfiltration of credentials and sensitive data. The vulnerability permits model-initiated file access by exploiting a self-whitelisting mechanism that fails to properly validate media parent directory paths. Authentication is required, but the flaw affects confidentiality with a CVSS score of 6.0.

Remote authenticated attackers can overwrite arbitrary files on OpenClaw servers by uploading malicious tar archives containing symbolic links to the SSH sandbox feature. The vulnerability allows escaping sandbox restrictions to modify critical system files, enabling potential remote code execution or denial of service. Affects OpenClaw versions before 2026.3.31. No public exploit identified at time of analysis, with EPSS data unavailable for this recent CVE.

OpenClaw versions 2026.2.19 before 2026.3.31 allow authenticated attackers to suppress legitimate webhook events across different accounts in multi-tenant deployments by exploiting improper cache isolation in the Zalo webhook replay-deduplication mechanism. An attacker with control of one authenticated Zalo webhook path can match event_name and message_id parameters to suppress events on victim accounts, causing denial of service to webhook processing. No public exploit code or active exploitation has been identified, though the vulnerability requires valid authentication and incremental exploitation (AT:P), limiting immediate risk.

Spring Boot applications configured with ApplicationPidFileWriter are vulnerable to local file corruption when a high-privilege user can write to the PID file directory. An attacker with high privileges and write access to the PID file location can corrupt arbitrary files each time the application restarts, achieving denial of service or data integrity violations. Exploitation requires local access and elevated privileges, limiting real-world impact to co-resident or insider threat scenarios. No active exploitation has been publicly reported.

{random.value}` property placeholder are generated with a non-cryptographic pseudo-random number generator, making them unsuitable for use as secrets such as passwords, tokens, or keys. The numeric `${random.int}` and `${random.long}` placeholders are even weaker, drawing from a small predictable range, while `${random.uuid}` is explicitly unaffected. There is no public exploit identified at time of analysis (SSVC exploitation: none; EPSS 0.03%), so the practical risk depends entirely on whether a given deployment actually used these placeholders to seed real secrets.

Spring Boot's Cassandra auto-configuration fails to verify hostnames during SSL/TLS connection establishment to Cassandra servers, enabling man-in-the-middle attackers on the local network to intercept credentials and data by presenting a valid certificate for any domain. Affects Spring Boot 2.7.0-4.0.5; vendor-released patches available for all supported versions (4.0.6, 3.5.14, 3.4.16, 3.3.19, 2.7.33). No public exploit code identified at time of analysis.

Local privilege escalation and session hijacking in Spring Boot allows attackers with local access to hijack authenticated sessions or execute arbitrary code by taking control of the ApplicationTemp directory. The vulnerability affects Spring Boot versions 2.7.0 through 4.0.5 when server.servlet.session.persistent is enabled, requiring attack persistence across application restarts. VMware has released patches for all supported branches (4.0.6, 3.5.14, 3.4.16, 3.3.19, 2.7.33), though unsupported versions remain vulnerable. No active exploitation confirmed at time of analysis.

KDE Dolphin before 25.12.3 allows sandboxed applications (running under Flatpak or AppArmor confinement) to bypass sandbox restrictions and open arbitrary files outside their containment boundary through the FileManager1 D-Bus protocol implementation. An attacker controlling a sandboxed application can exploit this to access sensitive files or execute scripts with user interaction, circumventing the intended isolation model.

When configured to use an SSL bundle, Spring Boot's RabbitMQ auto-configuration does not perform hostname verification when connecting to the RabbitMQ broker. Affected: Spring Boot 4.0.0-4.0.5 (fix 4.0.6), 3.5.0-3.5.13 (fix 3.5.14) per vendor advisory.

A vulnerability has been found in aligungr UERANSIM up to 3.2.7. The affected element is the function rls::DecodeRlsMessage in the library src/lib/rls/rls_pdu.cpp of the component Radio Link Simulation Layer. The manipulation of the argument pduLength leads to uncaught exception. The attack may be initiated remotely. Upgrading to version 3.2.8 is sufficient to fix this issue. The identifier of the patch is ca1a66fffe282767bb08618af9f848e3b68ea47b. It is suggested to upgrade the affected component. This behavior is related to CVE-2024-37877. The vendor was contacted early, responded in a very professional manner and quickly released a fixed version of the affected product.