Authenticated OS command injection in Coolify before 4.0.0-beta.471 lets any user holding destination management permissions execute arbitrary commands as root on managed servers. The 'network' parameter in the Destination Network Management feature is passed unsanitized into shell commands, yielding full root-level remote code execution on the host. No public exploit identified at time of analysis; the issue is fixed in 4.0.0-beta.471.
Authenticated remote code execution in Coolify (self-hosted PaaS) before 4.0.0-beta.470 lets a low-privileged authenticated user run arbitrary commands on the deployment host. The flaw is an OS command injection in the Nixpacks build pack: the user-supplied install_command build parameter is concatenated unsanitized into a shell command executed during the build phase, allowing escape from the build context to host-level command execution. There is no public exploit identified at time of analysis, and it is not listed in CISA KEV, but the path to abuse is straightforward for any user who can configure a deployment.
Remote code execution in FrontAccounting before 2.4.20 lets an authenticated user abuse the attachment upload handler to plant a PHP web shell. The handler fails to validate the unique_name parameter, so traversal sequences write attacker-controlled files outside the attachments directory and into the web root, and because file extensions are not validated, an uploaded PHP file executes as the web server user. Publicly available exploit code exists and a vendor patch is available, though the issue is not listed in CISA KEV.
WAF managed-rule body inspection on AWS Application Load Balancer (ALB) can be bypassed by remote actors who fragment an HTTP/2 request body across multiple frames so that only a partial body is inspected before reaching the backend. The flaw (CWE-444, HTTP request smuggling) affects only ALB target groups serving HTTP/2 traffic with AWS WAF enabled, and lets attackers slip malicious payloads past WAF managed rules. No public exploit identified at time of analysis and it is not on CISA KEV; AWS scores it 7.9 (CVSS 4.0) with impact falling on the protected backend rather than the ALB itself.
WAF inspection bypass in Amazon CloudFront (with AWS WAF enabled) lets remote actors smuggle malicious request bodies past managed rule inspection by fragmenting the HTTP/2 request body across frames so only a partial body is examined. The flaw (CWE-444, request smuggling) defeats the protective security control rather than CloudFront itself, allowing attacks the WAF would normally block to reach the protected origin. AWS remediated it server-side with no customer action required; there is no public exploit identified at time of analysis and it is not listed in CISA KEV.
OS command injection in the D-Link DCS-935L Wi-Fi network camera (firmware 1.10.01) lets a remote attacker inject arbitrary operating-system commands through the UID POST parameter handled by the sub_400E40 function in setconf.cgi. The CVSS 4.0 vector requires low privileges (PR:L), so an attacker needs some level of authenticated/session access, after which they gain full confidentiality, integrity, and availability impact on the device. Publicly available exploit code exists (disclosed via VulDB), though there is no public exploit identified as actively exploited in the wild.
DQL (Dgraph Query Language) injection in Dgraph's GraphQL-to-DQL query rewriter (versions <= v25.3.3) lets remote unauthenticated attackers inject arbitrary query blocks through the unauthenticated checkUserPassword GraphQL query. User-supplied password values are interpolated into a checkpwd() DQL call via fmt.Sprintf without escaping, so a double-quote breaks out of the string literal and appends attacker-controlled DQL. A working PoC exists (publicly available exploit code exists); the issue is not in CISA KEV and no active exploitation is reported, but server-side execution of injected blocks is confirmed, enabling schema/data enumeration and resource exhaustion.
The APCu Manager WordPress plugin before 4.5.0 does not escape APCu object-cache keys before rendering them in an admin-area page, leading to a Stored Cross-Site Scripting vulnerability. When a persistent object cache is enabled, cache keys derived from unsanitised user input (e.g. a transient name created by another APCu Manager WordPress plugin before 4.5.0 from an unauthenticated request) are output without escaping and execute arbitrary JavaScript in the session of an administrator viewing the page.
Stack-based buffer overflow in the Wavlink WL-NU516U1-A wireless range extender (firmware M16U1_V240425) lets a remote, low-privileged attacker corrupt memory by sending an oversized Guest_ssid POST parameter to the sub_407504 function in /cgi-bin/wireless.cgi. The flaw was reported by VulDB, publicly available exploit code exists, and it carries a CVSS 4.0 base score of 7.4; there is no evidence of active exploitation in CISA KEV at this time.
Stack-based buffer overflow in the Tenda JD12L router (firmware 16.03.53.23) lets remote attackers corrupt memory via the 'page' parameter handled by the fromNatStaticSetting function at the /goform/NatStaticSetting endpoint. Publicly available exploit code exists, and the CVSS 4.0 vector (PR:L) indicates an authenticated user on the device's web interface can trigger high confidentiality, integrity and availability impact. There is no public exploit identified as actively used in the wild - exploitation is proof-of-concept only at this time.
Stack-based buffer overflow in the Tenda JD12L router (firmware 16.03.53.23) lets a remote attacker corrupt memory by manipulating the 'page' parameter handled by the fromAddressNat function at the /goform/addressNat endpoint, potentially achieving code execution or device crash. Publicly available exploit code exists and the issue was disclosed by VulDB, though there is no public exploit identified as actively exploited in the wild. With a CVSS 4.0 base score of 7.4 and PR:L, exploitation requires some level of authenticated access to the web management interface.
Stack-based buffer overflow in the Tenda JD12L router (firmware 16.03.53.23) allows remote attackers to corrupt memory by manipulating the security_5g argument sent to the formWifiBasicSet handler at /goform/WifiBasicSet, potentially achieving denial of service or arbitrary code execution on the device. The flaw is network-reachable but, per the CVSS 4.0 vector, requires low-level authentication (PR:L), and publicly available exploit code exists. There is no public exploit identified as actively exploited in CISA KEV, and the EPSS score was not provided.
Buffer overflow in the Edimax EW-7478APC wireless access point (firmware 1.04) lets remote attackers corrupt memory by sending an oversized ShareName or SelectName argument to the formUSBFolder handler at /goform/formUSBFolder. The flaw is reachable over the network and publicly available exploit code exists, though no active exploitation has been reported. Per the CVSS 4.0 vector it requires low-level authentication (PR:L) but yields full confidentiality, integrity, and availability impact on the device.
Remote buffer overflow in the Edimax EW-7478APC wireless access point (firmware 1.04) lets an authenticated attacker corrupt memory through the formUSBAccount handler at /goform/formUSBAccount by supplying oversized UserName or Password values in a POST request, enabling likely code execution or device crash. The flaw is remotely reachable and publicly available exploit code exists (proof-of-concept, CVSS 4.0 base 7.4), though there is no public exploit identified as actively used in the wild. The vendor was notified but did not respond, so no fix is expected.
Buffer overflow in the Edimax EW-7478APC wireless access point (firmware 1.04) lets remote attackers overflow a buffer by manipulating the selSSID parameter in a POST request to the /goform/formQoS endpoint, potentially crashing the device or executing code on it. The flaw is network-reachable and publicly available exploit code exists (disclosed via VulDB and a Notion writeup), but it is not listed in CISA KEV and no active exploitation has been confirmed. The vendor was notified but did not respond, and no patched firmware has been published.
SQL injection in FrontAccounting's get_gl_transactions() function lets authenticated users holding the SA_GLANALYTIC permission read arbitrary database contents by injecting into the unparameterized filter_type parameter of a SQL IN() clause. The flaw enables boolean-based blind extraction of sensitive journal-entry and general-ledger data using reliable response-size differentials. Publicly available exploit code exists and a vendor patch shipped in 2.4.20, though there is no public exploit identified as actively exploited.
SQL injection in FrontAccounting's Audit Trail report handler (reporting/rep710.php) lets authenticated users holding the SA_GLANALYTIC permission run arbitrary SQL through the PARAM_2 and PARAM_3 POST parameters in all releases prior to 2.4.20. Attackers can extract arbitrary database contents via UNION-based injection or knock the application offline by amplifying SLEEP() across JOINed result sets to exhaust database connections. Publicly available exploit code exists (VulnCheck/Jiva Security writeup), but the issue is not in CISA KEV, so it is no public exploit identified as actively exploited.
SQL injection in FrontAccounting before 2.4.20 allows authenticated attackers to extract arbitrary database contents by injecting UNION SELECT payloads into the PARAM_0 POST parameter of the Bank Statement report handler (rep601.php). Attackers with valid low-privilege accounts can dump usernames, password hashes, and email addresses from the users table, with results rendered into the generated PDF report. Publicly available exploit code exists and a vendor patch was released in 2.4.20; the flaw is not listed in CISA KEV and no active exploitation is confirmed.
Root command injection in OpenWrt's LuCI web interface (luci-proto-openvpn through 0.11.1) lets an authenticated user with OpenVPN protocol configuration access run arbitrary commands as root. The flaw lives in the generateKey ubus method, where the cl_meta parameter is passed unescaped into a popen() shell call. No public exploit has been identified at time of analysis, but a vendor advisory (GHSA-pm9w-522m-8rrh) and fix commit exist, and the CVSS 4.0 score of 8.7 reflects full confidentiality, integrity, and availability impact on the affected device.
Arbitrary file overwrite in libzypp before 17.38.10 lets a remote attacker who controls a software repository plant crafted repository metadata (repomd/SUSE content files) with '../' location paths that escape the repo root, allowing files anywhere on the system to be overwritten when a victim adds and refreshes that repository. Because zypper-based refresh operations typically run as root, this can escalate to privilege escalation or denial of service. There is no public exploit identified at time of analysis and the issue is not listed in CISA KEV.
Arbitrary code execution in Snowflake CLI versions prior to 3.19 lets an attacker run code in the context of any developer who bundles or deploys an attacker-supplied Snowpark project. The flaw lives in the Snowpark annotation processor callback template, where untrusted project content is interpolated directly into generated Python code (CWE-94). No public exploit has been identified at time of analysis, but the attack is straightforward and high-impact (CVSS 8.8) given that it executes with the victim's local privileges; exploitation hinges on the victim running the bundling/deployment workflow against malicious content.
Memory corruption in Apple's WebKit browser engine (Safari and the system WebView on iOS, iPadOS, and macOS before 26.5.2) allows a remote attacker to corrupt process memory when a victim loads maliciously crafted web content. The flaw is a use-after-free (CWE-416) and carries a CVSS 8.8 with required user interaction; there is no public exploit identified at time of analysis and it is not listed in CISA KEV, though Apple reported and patched it in coordinated releases on 2026 advisories.
SQL injection in Snowflake CLI versions 1.2.2 through 3.18.x allows an attacker to execute unintended SQL statements within a victim's authenticated Snowflake session by planting crafted repository content, project configuration, manifest, or specification input. When a developer processes that attacker-controlled content through a vulnerable command path, the injected SQL runs with the victim's session privileges, enabling data theft, modification, or destruction up to that user's authorization level. There is no public exploit identified at time of analysis, exploitation is not confirmed in CISA KEV, and EPSS is low at 0.31% (23rd percentile), reflecting the user-interaction requirement.
Arbitrary VS Code command execution in the Red Hat vscode-java extension allows a malicious Java source file to embed hidden commands inside JavaDoc hover Markdown, so that a developer who simply clicks a crafted link in a hover popup triggers attacker-chosen commands that can escalate to full system compromise in trusted workspaces. The flaw stems from the extension rendering JavaDoc hovers as fully-trusted Markdown, and it also affects Red Hat OpenShift Dev Spaces, which bundles the extension. There is no public exploit identified at time of analysis and the issue is not listed in CISA KEV, but the network reach combined with only a single click of user interaction makes it high-impact.
Memory corruption via type confusion in Apple's WebKit browser engine allows attackers to corrupt memory by luring a victim to maliciously crafted web content, affecting Safari, iOS/iPadOS, and macOS before version 26.5.2. The flaw (CWE-843) is network-reachable but requires user interaction (visiting a page), and no public exploit has been identified at time of analysis. Apple has shipped patches across Safari 26.5.2, iOS/iPadOS 26.5.2, and macOS Tahoe 26.5.2.
Memory corruption via use-after-free in Apple's WebKit browser engine allows remote attackers to corrupt memory when a victim processes maliciously crafted web content on Safari, iOS/iPadOS, or macOS prior to 26.5.2. The CVSS 3.1 score of 8.8 reflects network-reachable exploitation requiring only that a user open a malicious page; no public exploit is identified at time of analysis and the issue is not listed in CISA KEV. Successful exploitation typically serves as the initial stage of a browser-based attack chain (often paired with a sandbox escape) to achieve code execution in the renderer.
Unauthenticated information disclosure in phpUploader before 2.0.2 lets remote attackers harvest the entire uploaded-files database table simply by loading any page of the application. The index model runs an unbounded SELECT and serializes the full result set into an inline JSON script block, leaking uploader IP addresses, Argon2ID key hashes, internal filenames, and SHA-256 fingerprints. No public exploit identified at time of analysis, but the data is exposed in plain HTML source on every page, making extraction trivial; CVSS 4.0 base is 8.7 (High).
Broken access control in the ErpSaleOrderController of ruoyi-vue-pro (through version 2026.05) lets authenticated users holding only shipment-level 'erp:sale-out' permissions perform full create, read, update, and delete operations on financially sensitive ERP sale orders. The controller enforces the wrong permission namespace ('erp:sale-out' instead of the intended 'erp:sale-order'), so low-privileged warehouse/shipment operators effectively gain order-management authority. No public exploit identified at time of analysis and the issue is not in CISA KEV; an upstream fix exists as commit 5d1fd70.
Remote code execution in SzafirHost (KIR's Szafir electronic-signature host application) arises from a ZIP/JAR parser-confusion flaw: signature verification reads the archive's Central Directory while extraction reads local file headers sequentially, letting an attacker who controls the served native-library archive smuggle an unsigned malicious DLL/SO/DYLIB past the signature check. Versions prior to 1.2.2 are affected; the injected library is written to the native temp directory and loaded, yielding code execution on the victim's machine. No public exploit identified at time of analysis, and the flaw is not listed in CISA KEV.
Local privilege escalation in the attr package's getfattr and setfattr utilities (versions before 2.6.0) allows an attacker who controls a pathname component to swap it for a symbolic link during directory hierarchy traversal, redirecting attribute operations to arbitrary files. When these utilities are invoked by a privileged process over an attacker-influenced path, the attacker can read or write extended attributes on files outside their authority, leading to escalation to higher privileges. The flaw was reported by VulnCheck; no public exploit has been identified at time of analysis, and it is not listed in CISA KEV.
Local privilege escalation in the Linux acl utilities (libacl) before version 2.4.0 arises because the pathname-based functions acl_get_file(), acl_set_file(), acl_extended_file(), and acl_delete_def_file() follow symbolic links during ACL read/write operations. An attacker who controls any component of a pathname processed by a privileged caller can substitute a symlink to redirect ACL operations to arbitrary files, manipulating access control lists outside their authority. No public exploit identified at time of analysis, and the issue is not in CISA KEV; risk is gated on a privileged process operating on attacker-influenced paths.
Unauthenticated attendee PII exposure in Hi.Events through 1.9.0 lets remote attackers retrieve full attendee lists — including names, emails, and personal information — by hitting the public check-in list endpoints, which treat the check-in list's short_id as the only access control. Beyond reading data, an attacker who knows or guesses a short_id can also create and delete check-in records without authentication. No public exploit identified at time of analysis; the issue was reported by VulnCheck and is tracked via an upstream GitHub fix.
SQL injection in SigNoz observability platform (versions through 0.130.1) lets authenticated, low-privileged users inject arbitrary ClickHouse queries through the unsanitized rule ID path parameter of the alert-history endpoints. By URL-encoding quote characters that break out of the interpolated rule ID, an attacker can read every stored trace, log, and metric, and can abuse ClickHouse's url() table function to pivot into server-side request forgery against internal services. This is CWE-89 with a documented SSRF secondary impact; no public exploit identified at time of analysis and it is not listed in CISA KEV.
Improper authorization in the Maintenance Utility of Hitachi Virtual Storage Platform (VSP) enterprise storage arrays allows an authenticated low-privileged user to perform actions beyond their authorization, undermining the integrity and availability of the storage controller. The flaw affects the VSP E-series, VSP 5000-series, and VSP G/F-series families running pre-fix DKCMAIN and GUM firmware, and was reported by Hitachi itself; there is no public exploit identified at time of analysis and the issue is not listed in CISA KEV.
Promo code usage-limit bypass in Hi.Events through version 1.9.0 lets remote attackers redeem restricted/limited promo codes an unlimited number of times by exploiting a time-of-check/time-of-use gap between synchronous reservation validation and the asynchronous UpdateEventStatisticsJob that increments the usage counter. Because validation reads order_usage_count before the background job updates it, an attacker can sequentially reserve and complete many discounted orders - no concurrent/racing requests are even required. There is no public exploit identified at time of analysis and the issue is not listed in CISA KEV; the CVSS 4.0 base score is 8.2, reflecting high integrity impact on the platform's discount/pricing controls.
Unauthenticated denial-of-service in MDEx, an Elixir/Erlang CommonMark parsing library, lets an attacker crash the entire BEAM virtual machine by feeding a crafted JSON document to MDEx.parse_document/2's {:json, ...} source. Because each unique node_type value is interned as a permanent BEAM atom that is never garbage collected, a single deeply nested document can mint hundreds of thousands of atoms and exhaust the ~1,048,576-atom table, aborting every process on the node. The flaw affects mdex 0.4.3 through 0.13.1, is fixed in 0.13.2, and has no public exploit identified at time of analysis.
The issue was addressed with improved checks. This issue is fixed in Safari 26.5.2, iOS 26.5.2 and iPadOS 26.5.2, macOS Tahoe 26.5.2. A malicious website may exfiltrate data cross-origin.
SQL injection in Snowflake CLI versions 1.1.0 through 3.18.x (fixed in 3.19) lets crafted parameter values reach vulnerable command paths and execute unintended SQL within the user's active Snowflake session. An authenticated CLI user who is fed malicious input - via social engineering, a poisoned repository configuration, or compromised automation - can have arbitrary statements run against their session, with impact bounded by that session's privileges. No public exploit identified at time of analysis; EPSS is very low (0.11%, 1st percentile) and the SSVC exploitation status is 'none', so this is a not-yet-exploited but high-technical-impact issue.
Authenticated root command injection in the OpenWrt LuCI community Tailscale app (luci-app-tailscale-community) lets a logged-in web UI user run arbitrary OS commands as root through the tailscale.do_login RPC method. Because the user-supplied loginserver and loginserver_authkey values are placed inside a double-quoted shell string, $() and backtick substitutions are evaluated by the outer shell, turning configuration input into command execution. Reported by VulnCheck with a published GHSA advisory; no public exploit identified at time of analysis and the issue is not in CISA KEV.
Sandbox escape and code execution in Anthropic's Claude Code CLI (versions 2.1.38 through 2.1.162) lets a malicious repository break out of the macOS seatbelt sandbox by abusing git worktree handling. By creating a worktree named ".git" and combining symlink manipulation with git fsmonitor execution, an attacker can overwrite files in the user's home directory such as .zshenv to achieve code execution outside the sandbox. There is no public exploit identified at time of analysis and the issue is not in CISA KEV, but exploitation is realistic since it only requires a user to clone and open a booby-trapped repo; it is fixed in 2.1.163.
Local privilege escalation and kernel memory corruption in Apple iOS, iPadOS (before 26.5.2) and macOS Tahoe (before 26.5.2) allows a malicious or compromised app to write kernel memory or force unexpected system termination by supplying improperly validated input to a privileged interface. Apple resolved the flaw through improved input sanitization. There is no public exploit identified at time of analysis, EPSS is low (0.18%, 8th percentile), and CISA SSVC records exploitation status as none, though it rates the technical impact as total and considers it automatable.
Local privilege escalation in Matrix42 Empirum (versions before 25.5 and 26.x before 26.2) lets any authenticated low-privileged user gain SYSTEM execution by abusing the PBackupVSS.exe service's named pipe. The pipe \\.\pipe\PBackupVSS is created with an overly permissive DACL granting GENERIC_READ/GENERIC_WRITE to all authenticated users, so an attacker can send crafted IPC messages that drive the SYSTEM-level service into running an attacker-supplied shadow.exe from a controlled working directory. No public exploit identified at time of analysis and the issue is not listed in CISA KEV, but exploitation is mechanically straightforward once on the host.
Server-side request forgery in Nitter's /video media proxy endpoint lets unauthenticated remote attackers coerce the server into fetching arbitrary URLs and returning their HTTP responses, including AWS/GCP cloud metadata endpoints and internal-only network resources. The flaw stems from two compounding errors: the proxy never validates that target URLs belong to Twitter/X domains, and it ships with a hardcoded default HMAC signing key, so an attacker can locally compute the valid HMAC signature required for any URL of their choosing. No public exploit has been identified at time of analysis, but the technical details and the fixing commit are public.
{pk}/password/ endpoint, which fails to enforce object-level authorization (CWE-639 IDOR). By simply supplying another user's primary key, a domain-scoped admin can take over the highest-privileged account and gain full control of the mail platform. There is no public exploit identified at time of analysis and it is not listed in CISA KEV, but the fix is upstream in PR #4038 (commit a1878c4).
Cross-team data exposure in Coolify (self-hostable PaaS) before 4.0.0-beta.471 allows any authenticated user to read servers and projects owned by other teams by supplying their object IDs directly. The flaw is a broken object-level authorization (IDOR) where lookups are never scoped to the requester's team, breaching the multi-tenant boundary. No public exploit identified at time of analysis, and the issue is resolved in 4.0.0-beta.471.
Path traversal in the H-VUE web management subsystem of Gigamon GigaVUE-OS (GVOS) versions 5.16.1 and earlier allows remote attackers to read arbitrary files outside the intended web root, exposing sensitive system data. The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N) indicates network-reachable, low-complexity, unauthenticated exploitation with high confidentiality impact but no integrity or availability effect. A researcher GitHub repository tracking this CVE suggests proof-of-concept material may exist, but there is no public exploit explicitly confirmed and no CISA KEV listing at time of analysis.
Arbitrary file read in the Zephyr RTOS HTTP server (subsys/net/lib/http) lets an unauthenticated remote client retrieve any readable file on the mounted filesystem volume by abusing path-traversal sequences against a registered static-filesystem resource. Affecting Zephyr v4.0.0 through v4.4.0 with CONFIG_FILE_SYSTEM enabled and a static-FS resource registered, the flaw stems from the raw request path being concatenated to the web root without canonicalization in both the HTTP/1 and HTTP/2 front-ends. No public exploit identified at time of analysis, but the bug is trivial to trigger (CVSS 7.5, AV:N/AC:L/PR:N/UI:N) and a vendor patch is available.
Session hijacking in Pinpoint (open-source APM) through version 3.1.0 stems from the pinpointJwt session cookie being issued without the HttpOnly and Secure flags, letting client-side JavaScript read it via document.cookie and allowing it to traverse cleartext HTTP. An attacker who can land a stored or reflected XSS payload, or who can sniff network traffic, can steal the JWT session token and impersonate the victim. There is no public exploit identified at time of analysis and the issue is not on CISA KEV; the CVSS 4.0 base score is 7.6.
Uncontrolled memory consumption in the Perl module JavaScript::Minifier::XS before version 0.16 lets remote attackers exhaust a long-lived process's memory by repeatedly triggering minify() calls. The XS cleanup code frees only the NodeSet structures while leaking every per-token contents buffer (and the entire NodeSet on the empty-list early-return paths), so each invocation leaks heap memory until the process is OOM-killed. CISA's SSVC marks exploitation as automatable with partial technical impact; no public exploit has been identified and it is not in CISA KEV, though the leak is trivially reachable in any service that minifies attacker-supplied JavaScript.
Denial of service in Apache ActiveMQ (versions before 5.19.8, and 6.0.0 before 6.2.7) lets an authenticated user crash the broker by sending a crafted OpenWire message whose property map declares an excessively large encoded size. Because the map is unmarshaled without validating the declared size against actual payload bounds, the broker pre-allocates massive memory and hits an OutOfMemory condition. No public exploit identified at time of analysis and the issue is not in CISA KEV, but the low attack complexity and availability of a vendor patch make it a practical operational concern for exposed brokers.