Stored cross-site scripting in MCP Registry's catalogue UI allows any user with a publish token to inject arbitrary event handlers via the `websiteUrl` field by breaking out of an `href` attribute with an unescaped double-quote character. The server-side URL validator accepts quotes and the client-side `escapeHtml` helper fails to encode them in attribute context, enabling attackers to execute JavaScript on the registry.modelcontextprotocol.io origin with access to localStorage, XHR, and auth tokens. Vendor-released patch version 1.7.7 available; actively confirmed via proof-of-concept.
Open redirect vulnerability in Kargo UI OIDC login flow allows unauthenticated remote attackers to redirect users to arbitrary external websites via a malicious redirectTo query parameter. Versions prior to 1.7.10, 1.8.13, 1.9.8, and 1.10.2 are affected. This requires user interaction (clicking a crafted link) but can facilitate phishing attacks by making malicious redirects appear legitimate within the Kargo authentication flow.
Cross-site scripting in wlc command-line client versions prior to 2.0.0 allows authenticated users with high privileges to inject malicious HTML/JavaScript into API responses, which are then embedded unescaped in HTML output. When the HTML output is rendered in a browser, this enables XSS attacks. The vulnerability requires explicit use of the HTML output format (non-default), user interaction to open/view the HTML file, and elevated API credentials, limiting real-world risk despite the network vector.
Availability degradation in Bouncy Castle BC-FJA cryptographic library versions 2.1.0 through 2.1.2 on Linux X86_64 systems with AVX or AVX-512f SIMD extensions affects GCM-128 and GCM-512 operations, allowing local attackers to cause denial of service without authentication. The vulnerability is associated with optimized assembly implementations (gcm128w, gcm512w) and results in availability impact with no confidentiality or integrity compromise. No public exploit code or active exploitation has been identified at time of analysis.
Mass assignment vulnerability in Open WebUI's folder creation endpoint allows authenticated attackers to create folders in other users' accounts by exploiting Pydantic's extra='allow' configuration. An attacker with a valid account can supply an arbitrary user_id in the POST request body, overwriting the server-assigned value and persisting folders under a victim's account without their knowledge. The attacker can use this to plant phishing content, spam folders, or degrade user experience for targeted victims.
External Secrets Operator versions 0.1.0 through 2.4.0 allow authenticated users with ExternalSecret creation permissions to escalate privileges by crafting Service Account token templates that cause the operator to generate long-lived tokens for any service account in the namespace. An attacker can impersonate service accounts without requiring direct TokenRequest or Secret creation permissions, effectively bypassing RBAC controls. The attack requires the attacker already has ExternalSecret creation permissions and the cluster must have service account token generation enabled, limiting the practical scope to already-privileged users seeking lateral privilege expansion within a namespace.
Authenticated administrators in Flarum can read arbitrary files and trigger server-side request forgery via LESS injection in theme color settings. The vulnerability exploits an incomplete patch for CVE-2023-27577 that restricted @import and data-uri() only in the custom_less setting but failed to apply the same restrictions to other LESS config variables such as theme_primary_color and theme_secondary_color. An attacker with admin credentials can inject arbitrary @import directives into compiled forum.css, exposing sensitive files or making outbound HTTP requests to internal networks and cloud metadata endpoints. Vendor-released patches: Flarum 1.8.16 and 2.0.0-rc.1.
Stored XSS in Open WebUI's AccountPending.svelte allows an authenticated admin to inject arbitrary JavaScript into the pending user overlay content, which executes in the browser session of any user holding the 'pending' role. The flaw stems from invoking DOMPurify before rather than after marked.parse(), permitting Markdown link syntax with javascript: URIs to survive sanitization and render as executable HTML via Svelte's {@html} directive. No CISA KEV listing exists, but a proof-of-concept is confirmed by SSVC data; EPSS is low at 0.03% (8th percentile), consistent with the high-privilege prerequisite limiting mass exploitation.
go-git versions prior to 5.18.0 and 6.0.0-alpha.2 leak HTTP authentication credentials when following cross-host redirects during smart-HTTP clone and fetch operations. Remote unauthenticated attackers controlling a redirect target can capture credentials intended for the original repository host. User interaction (initiating a clone/fetch to a malicious or compromised server) is required. Vendor-released patches are available in v5.18.0 and v6.0.0-alpha.2.
Race condition in the Linux kernel nvme-pci driver's nvme_poll_irqdisable() function causes an unbalanced IRQ enable/disable pair that crashes the kernel with a warning. Affected kernels from 5.7 through multiple stable branches are vulnerable when running PCIe NVMe storage with MSI-X interrupts: a concurrent NVMe device reset can change the IRQ vector between the disable_irq() and enable_irq() calls, making the kernel operate on different IRQ numbers. No public exploit identified at time of analysis and EPSS of 0.02% confirm this is a reliability/stability concern patched in kernel stable releases 6.1.167, 6.6.130, 6.12.78, 6.18.19, 6.19.9, and 7.0.
Race condition in the Linux kernel cgroup subsystem's task iterator exposes local low-privileged users to a denial-of-service condition when task migration and cgroup iteration execute concurrently. The cgroup infrastructure fails to advance active css_task_iters before a task is unlinked from cset->tasks during migration, allowing iterators to reference the wrong linked list and silently skip tasks - or in worst-case scenarios, cause css_task_iter_advance() to crash or loop infinitely on the destination css_set. No public exploit identified at time of analysis; EPSS of 0.02% at the 7th percentile reflects extremely low observed exploitation probability and aligns with the narrow race window required.
Race condition in the Linux kernel's yurex USB driver probe function allows a local low-privileged attacker to cause a denial of service by triggering a timing window between URB submission and bbu member initialization. Affected are all kernel versions from the initial commit through the stable branch fix points (patched in 5.10.253, 5.15.203, 6.1.167, 6.6.130, 6.12.78, 6.18.19, 6.19.9, and 7.0). No public exploit exists and the issue is not listed in CISA KEV; EPSS of 0.02% (7th percentile) reflects negligible widespread exploitation probability.
Race condition in the Linux kernel's USB RNDIS gadget function driver (f_rndis) allows a local low-privileged attacker to crash the kernel by concurrently manipulating class/subclass/protocol configfs attributes without mutex protection. Identified during code inspection - not observed in active exploitation - this vulnerability affects multiple stable kernel branches from 4.14 through 7.0-rc3, with patches released across all maintained stable series. With an EPSS of 0.02% (7th percentile), no public exploit, and no CISA KEV listing, real-world risk is low but meaningful on embedded or IoT devices using Linux as a USB RNDIS peripheral.
Kernel panic triggered by a race condition in the UFS Host Controller Driver (ufshcd) during system suspend affects Linux systems using Universal Flash Storage hardware where UFSHCD_CAP_CLK_GATING is not supported. The flaw allows a local low-privileged user - or automated power management - to crash the kernel by triggering a suspend sequence while ufshcd_rtc_work() is concurrently executing, producing an ARM64 asynchronous SError interrupt that halts the system. No public exploit code exists and no active exploitation has been identified; with an EPSS of 0.02% this is a low-probability but confirmed-availability-destroying defect patched across multiple stable kernel branches.
OS command injection in Vim's ':find' tab-completion mechanism allows an attacker who controls a file's contents to execute arbitrary shell commands under the victim's user account, by exploiting the missing P_SECURE flag on the 'path' option prior to version 9.2.0435. Because 'path' lacked this flag, a file modeline could inject backtick-wrapped shell commands that Vim silently executed during file-name completion. No active exploitation has been confirmed (SSVC: Exploitation none, EPSS 0.04%), and the vendor-released patch is available in Vim 9.2.0435.
OS command injection in Vim's netrw plugin prior to version 9.2.0383 allows arbitrary shell command execution at the privilege level of the Vim process when a user is induced to open a crafted sftp:// or file:// URL. The root cause is a two-part failure: the temporary file suffix extracted from the remote filename was not restricted to safe characters, and the resulting tmpfile path was passed unescaped to shell command construction in netrw#os#Execute(). No public exploit has been identified at time of analysis and SSVC confirms no observed exploitation, but the commit diff includes a regression test that demonstrates the injection path directly.
n8n-MCP prior to version 2.47.13 logs sensitive credential material from authenticated MCP tool-call requests when running in HTTP transport mode, allowing disclosure of bearer tokens, OAuth credentials, API keys, and webhook authentication headers to any system with access to server logs. The vulnerability requires valid authentication (AUTH_TOKEN) and affects deployments where logs are collected, forwarded to external systems, or viewable outside the request trust boundary; no public exploit code has been identified.
Onyx versions before 3.0.9, 3.1.6, and 3.2.6 permit authenticated users to terminate any other user's active chat session via the POST /chat/stop-chat-session/{chat_session_id} endpoint without verifying session ownership. An attacker with valid credentials can interrupt another user's LLM generation mid-stream by supplying a known session UUID, causing denial of service to targeted chat sessions. Vendor-released patches are available, and no public exploit code or active exploitation has been identified at time of analysis.
Nil-pointer dereference in free5GC's Unified Data Repository (UDR) v4.2.1 allows an OAuth2-authenticated network attacker to repeatedly panic the `amf-subscriptions` DELETE handler, forcing HTTP 500 responses and elevated CPU and log overhead on each request. The flaw stems from a missing `return` after the subscription-not-found branch in `RemoveAmfSubscriptionsInfoProcedure`, causing the handler to dereference a nil map entry even when it has already set a 404 problem-details response. No public exploit is in CISA KEV, but a full end-to-end proof-of-concept is embedded in GHSA-4rqf-grm6-vf75 and was validated against the official Docker compose lab, making exploitation straightforward for any holder of a valid `nudr-dr` token.
Improper permission enforcement in Wagtail CMS allows authenticated low-privilege users to access the page history report for pages they lack edit rights on, resulting in potential disclosure of sensitive revision content. Affected versions are pip/wagtail prior to 7.0.7 and 7.1 through below 7.3.2. No public exploit code exists and no active exploitation has been identified; EPSS of 0.03% (7th percentile) and CISA SSVC classification of 'Exploitation: none' confirm this as a low-priority disclosure issue limited to the authenticated CMS user population.
Open WebUI versions up to 0.8.12 allow authenticated users to enumerate members of private standard channels via the GET /api/v1/channels/{id}/members endpoint, which lacks access control checks present on other channel endpoints. An attacker who knows a private channel's UUID can retrieve the full list of members including their names, emails, roles, and profile images, enabling targeted social engineering and organizational structure reconnaissance. The vulnerability is fixed in version 0.9.0.
Open WebUI versions up to 0.8.12 allow authenticated users to enumerate all knowledge bases across the instance via an incomplete access control allowlist in the retrieval collection validation function. The `_validate_collection_access` function only enforces ownership checks for collections matching `user-memory-*` and `file-*` patterns, allowing any authenticated user to directly query the system-level `knowledge-bases` meta-collection and retrieve the IDs, names, and descriptions of every knowledge base regardless of ownership. This information disclosure vulnerability serves as an enabler for subsequent attacks including knowledge base destruction and content injection, transforming these attacks from theoretically exploitable (requiring random UUID guessing) to trivially exploitable (UUIDs enumerable). CVSS score 4.3 (network-accessible, low privilege required, low confidentiality impact). Patched in version 0.9.0.
Bugsink versions 2.1.2 and earlier contain a webhook URL validation bypass (SSRF) where malformed URLs with backslashes and @ symbols pass validation checks but are interpreted differently by Python's urllib parser versus the requests HTTP client, allowing attackers with webhook configuration access to direct outbound POST requests to blocked hosts including loopback and private addresses. The vulnerability is narrower than full SSRF because requests do not follow redirects, the request shape is constrained by URL normalization, and this only affects webhook integrations, not arbitrary outbound proxying.
The text-to-speech endpoint in AnythingLLM prior to v1.12.1 exposes an Insecure Direct Object Reference (IDOR) allowing any authenticated workspace member to retrieve audio renderings of another user's private chat responses by supplying a foreign chatId. The route validated workspace membership but omitted per-row ownership enforcement, making lateral access within a shared workspace trivial for a low-privileged attacker. No active exploitation is confirmed (absent from CISA KEV), though SSVC classifies exploitation as poc-level, and EPSS is negligible at 0.01%, consistent with a targeted rather than opportunistic threat.
Kimai versions 2.32.0 through 2.55.x allow System-Admin users with invoice template upload permission to read arbitrary files from the PHP server via malicious PDF invoice templates. The vulnerability exploits mPDF's SetAssociatedFiles() function combined with unsanitized Twig template rendering to access any file readable by the PHP worker process and embed it within generated PDF invoices. No public exploit code or active exploitation has been identified; patch available in version 2.56.0.
Privilege escalation in SysReptor prior to version 2026.29 allows authenticated User Admins to hijack Superuser accounts by exploiting an overly permissive email-change capability combined with the optional Forgot Password reset flow. When the 'Forgot Password' feature is enabled (non-default) and the targeted Superuser has no MFA configured, a User Admin can redirect the password reset token to an attacker-controlled address and authenticate as Superuser, gaining access to the Django backend (/admin) and full installation settings control. No public exploit exists and EPSS stands at 0.02% (5th percentile), consistent with SSVC's 'none' exploitation status, but the impact when conditions are met is significant for the affected SysReptor installation.
Open redirect in draw.io's GitLab OAuth flow allows an attacker to hijack the authorization popup by supplying a crafted `?gitlab=` URL parameter, replacing the legitimate GitLab server with an attacker-controlled host. All draw.io versions prior to 29.7.9 are affected; users who follow a crafted link and click through the 'Authorize in GitLab' dialog expose their credentials and OAuth state tokens to the attacker's phishing endpoint. No public exploit has been identified and CISA has not listed this in KEV, but the social-engineering barrier is low given draw.io's shareable-URL model.
Dell PowerScale OneFS versions 9.5.0.0 through 9.12.0.1 contain an insufficient logging vulnerability that allows low-privileged local attackers to tamper with information without generating adequate audit trails, enabling attack obfuscation and compliance violation. The vulnerability affects multiple version branches across OneFS 9.5 through 9.12, with no public exploit code identified at time of analysis. CVSS score of 3.3 reflects low-to-medium integrity impact with local access requirement and low complexity.
Server-side template injection in OpenStack Ironic through version 35.x allows authenticated administrators to disclose sensitive information by rendering unsandboxed Jinja2 templates in the instance_info['ks_template'] parameter. The vulnerability requires high-privilege user interaction and has low confidentiality impact with no integrity or availability consequences.
The EqualsUri function in uriparser before version 1.0.2 incorrectly classifies structurally distinct URIs as equivalent due to flawed absolutePath comparison logic when a host is present. An attacker can craft two different URIs that the library treats as identical, potentially bypassing URI-based access control checks or authentication mechanisms that rely on URI comparison. The vulnerability affects all versions before 1.0.2 and requires local access with high attack complexity; the impact is limited to integrity (logic bypass) with no confidentiality or availability impact.
Pointer difference truncation to signed int in uriparser before version 1.0.2 allows local attackers to cause integer overflow and data integrity issues through specially crafted URI inputs. The vulnerability stems from unsafe casting of pointer arithmetic results (afterLast - first) to int, which can overflow on systems where pointer differences exceed INT_MAX, leading to buffer overflows, incorrect memory calculations, and potential information disclosure. While CVSS score is low (2.9) due to local attack vector and high complexity, the fix adds comprehensive overflow detection using SIZE_MAX checks, indicating real risk in applications processing untrusted URIs locally.
Reflected cross-site scripting (XSS) in absinthe_plug GraphiQL interface allows unauthenticated attackers to execute arbitrary JavaScript in victims' browsers via the query GET parameter. The vulnerability exploits incomplete input escaping in the js_escape/1 function, which fails to escape backslashes before embedding user-controlled query strings into inline JavaScript. An attacker can bypass existing single-quote and newline escaping by prefixing a quote with a backslash (e.g., \'), breaking out of the string context. Vendor-released patch available (version 1.10.2 and later); exploitation requires user interaction (clicking a malicious link).
Server-side request forgery (SSRF) in FastGPT prior to version 4.14.17 allows authenticated users with App editing privileges to bypass SSRF protections in the lafModule workflow node's fetchData function, enabling arbitrary HTTP requests to internal and private network addresses via unvalidated user-controlled URLs passed to axios without filtering against the application's isInternalAddress blocklist.
SQL injection vulnerability in pgx (Go PostgreSQL driver) prior to version 5.9.2 allows authenticated attackers to manipulate queries when the non-default simple protocol is used in conjunction with dollar-quoted string literals containing attacker-controlled placeholder-like text. The vulnerability requires specific configuration (simple protocol mode enabled) and precise SQL structure (dollar-quoted strings with embedded placeholder syntax), making exploitation unlikely in typical deployments but possible in applications explicitly using QueryExecModeSimpleProtocol.
MCP Registry's GitHub OIDC token exchange allows cross-registry replay attacks due to use of a shared global audience string instead of registry-specific identifiers. An attacker controlling or observing any registry deployment can capture a legitimately issued OIDC token and replay it to another registry instance sharing the same codebase to obtain publish-capable JWTs for the victim GitHub owner namespace, breaking deployment isolation. The vulnerability affects all versions prior to 1.7.6; vendor-released patch available.
Reflected cross-site scripting (XSS) in SourceCodester Pizzafy Ecommerce System 1.0 via the page parameter in /admin/index.php allows remote attackers to inject malicious scripts that execute in a victim's browser with user interaction. CVSS 2.1 (low) reflects the requirement for user click-through; however, the vulnerability is disclosed publicly with proof-of-concept code available.
AsusPTPFilter driver allows local authenticated users to bypass security mechanisms via crafted IOCTL requests, potentially leaking restricted touchpad information or disabling the touchpad entirely. The vulnerability requires local access and low-level privileges but impacts the integrity and availability of the touchpad subsystem. CVSS 2.0 reflects limited scope (low severity across confidentiality, integrity, availability), but the attack vector is local and requires existing user privileges.
Unrestricted PHP file upload via the plugin management interface in Emlog prior to version 2.6.11 enables an attacker to upload and execute arbitrary PHP code, resulting in complete server compromise and persistent backdoor installation. All Emlog deployments below 2.6.11 are affected. No public exploit has been identified at time of analysis, and EPSS places exploitation probability at 0.04% (13th percentile); however, successful exploitation carries critical technical impact - full control of the underlying server.
Open redirect vulnerability in MCP Registry TrailingSlashMiddleware allows remote attackers to craft protocol-relative URLs that bypass path validation, redirecting users from the trusted registry domain to attacker-controlled sites. Affected versions 1.1.0 through 1.7.4 are vulnerable; vendor-released patch available in version 1.7.5. No public exploit code exists at time of analysis, but the vulnerability is trivially exploitable via simple HTTP requests without authentication.