h2oGPT through 0.2.1 contains a path traversal vulnerability in the OpenAI-compatible files API that allows unauthenticated remote attackers to read, write, and delete arbitrary files accessible to the server process by supplying traversal sequences in the bearer token. The get_user_dir function in openai_server/backend_utils.py uses the bearer token string unsanitized as a path component via os.path.join, and because the default API key is EMPTY authentication is bypassed, enabling attackers to traverse outside the intended user directory through the file content, delete, and upload endpoints to achieve remote code execution by writing to startup hooks or application-loaded files.
AgentGPT through 1.0.0 contains an authorization bypass through user-controlled key vulnerability that allows authenticated users to attach tasks to another user's agent run by supplying a target run_id in the request body without ownership verification. The AgentCRUD.create_task and validate_task_count functions look up the target AgentRun using the client-supplied run_id without confirming the run belongs to the requesting user, enabling an attacker who obtains a valid run_id to corrupt task history, exhaust the per-run loop budget, and drive LLM costs against the victim's run.
APIFold reads an OpenAPI 3.x or Swagger 2.x specification and generates a live, production-ready MCP server endpoint. Prior to commit 7f19b52280f414f57af2b79a95333d1c8fbeece5, the `/webhooks/:serverSlug/:eventName` endpoint accepts arbitrary unauthenticated JSON and stores it in Redis and the `webhook_events` PostgreSQL table without any signature check or authentication requirement. The root cause is that `createWebhookRouter` is called at `server.ts:188` without a `validators` map, so `receivers.ts:80`'s optional-chaining guard evaluates to `undefined` and the signature-validation block (`receiver.ts:81–95`) is unconditionally skipped. Any unauthenticated network client that knows a valid server slug can inject arbitrary payloads, which are subsequently served as trusted resource state to legitimate MCP clients. Commit 7f19b52280f414f57af2b79a95333d1c8fbeece5 patches the issue.
ITFlow provides an IT documentation, ticketing and accounting system for small managed service providers. Prior to version 26.05, low-privileged authenticated agent can retrieve plaintext credentials and TOTP secrets belonging to another client by directly requesting the credential edit modal with an arbitrary `credential_id`. The endpoint does not enforce client scoping or object-level authorization before loading and decrypting the credential record. Version 26.05 fixes the issue.
Tugtainer is a self-hosted app for automating updates of Docker containers. Versions prior to 1.30.2 are vulnerable to Server-Side Template Injection (SSTI) in the notification template feature. The `title_template` and `body_template` fields are rendered using an unsandboxed `jinja2.Environment`, allowing any authenticated user to execute arbitrary OS commands as root inside the container. Version 1.30.2 fixes the issue.
Shopper is a Headless e-commerce Admin Panel. Prior to 2.8.0, three related defects on admin Livewire components allowed data tampering, sensitive data disclosure, and stored XSS. First, several Livewire components in the admin panel exposed Eloquent model identifiers as public properties without the `#[Locked]` attribute. An authenticated user could rewrite the wire payload from the browser to target any record id, bypassing the implicit scoping enforced by the page routing. Second, `Customers/Create::store()` re-passed a `Hidden` `_password` form field straight into the create payload. The plaintext password was rendered into the HTML and transported through the Livewire snapshot in clear text, exposing credentials in the page DOM and in any logging that captures Livewire payloads. Finally, the product barcode field was rendered through `DNS1DFacade::getBarcodeHTML()` with `{!! !!}`. An attacker with `edit_products` permission could persist malicious payload in the barcode field that would execute in the browser of any admin user viewing that product, enabling session theft and privileged-action chaining. Starting in v2.8.0, all vulnerable Livewire model identifiers are now marked `#[Locked]`; `Customers/Create` no longer round-trips the password through a Hidden form field; the plaintext password is hashed at action boundary and never returned to the client; and the product barcode rendering now escapes the value before passing it to the barcode generator and the output is wrapped in an `<svg>` context that does not interpret event handlers. No known workarounds are available.
DbGate is cross-platform database manager. In versions 7.1.8 and prior, DbGate's JSON script runner (`POST /runners/start`) allows remote code execution via code injection in the `functionName` parameter of JSON script `assign` commands. The `functionName` value is interpolated directly into dynamically generated JavaScript source code via string concatenation. The generated code is then executed in a forked Node.js child process. Version 7.1.9 contains a patch.
Kata Containers is an open source project focusing on a standard implementation of lightweight Virtual Machines (VMs) that perform like containers. Versions prior to 3.31.0 ship with a default configuration that allows pod creators to inject arbitrary command-line arguments into the virtiofsd process through the `io.katacontainers.config.hypervisor.virtio_fs_extra_args` pod annotation. By injecting `-o source=/` along with `--no-announce-submounts` and `--sandbox=none`, an attacker can override the virtiofsd shared directory to serve the entire host root filesystem into the guest VM. Combined with the `kernel_params` annotation (also enabled by default) to activate the agent debug console, the attacker can mount the host filesystem from inside the VM and read or write any file on the host, including /etc/shadow. Version 3.31.0 patches the issue.
Joomla Extension - joomshaper.com - Unauthenticated SQL injection in Easy Store extension 1.0.0-2.0.1 - Improper validation of order parameters lead to an unauthenticated SQL injection in easystore, allowing full DB read access including credentials and sessions.
Joomla Extension - joomshaper.com - cross-customer order and personal information disclosure in Easy Store extension 1.0.0-2.0.1 - Improper access checks allow logged in users to retreive order and customer information of any order in the system.
Joomla Extension - joomshaper.com - unauthenticated payment/order forgery in Easy Store extension 1.0.0-2.0.1 - Critical order and payment information, including states, are processed from client side input, enabling unauthenticated attackers to manipulate payment and order states of arbritrary orders.
Void through 1.3.4 contains a path traversal vulnerability in the AI agent file-reading tools that allows network-adjacent attackers to read arbitrary host files outside the open workspace by injecting instructions into content the agent processes. Attackers can supply absolute paths or file:// URIs through the read_file, ls_dir, get_dir_tree, and search_* tools, which lack workspace confinement and bypass the approval gate, enabling silent exfiltration of sensitive files such as SSH private keys or cloud credentials via subsequent tool calls.
Fathom Lite through 1.3.1 contains a stored cross-site scripting vulnerability in the analytics collection endpoint that allows unauthenticated attackers to inject a javascript: URI into the Top Pages dashboard by supplying a crafted hostname and pathname to the unauthenticated /collect endpoint. The parseHostname and parsePathname functions perform no URI scheme validation, allowing a javascript: hostname combined with a newline-prefixed pathname to be stored and later rendered as an anchor href in the authenticated dashboard without sanitization, enabling session hijacking and full account takeover when an operator clicks the poisoned entry.
Overseerr through 1.35.0 contains an authorization bypass through user-controlled key vulnerability in the push subscription API that allows authenticated users to list, read, and delete any other user's push subscriptions by supplying an arbitrary userId in the path parameters. Attackers can exploit the missing ownership check in the affected handlers to access target user records without the filteredFields filter, leaking sensitive data including email addresses and plexId values.
Office-Word-MCP-Server through 1.1.11 contains a path traversal vulnerability in its document tools that allows attackers who can influence the filename argument to read arbitrary .docx files or create and overwrite .docx files outside the intended working directory. Attackers can supply absolute paths or ../ traversal sequences directly to document open and save operations, bypassing the check_file_writeable and ensure_docx_extension helpers which perform no base-directory confinement or realpath validation.
Proxygen lacked a generalized slow-consumer detection mechanism in its core HTTP session layer. A remote, unauthenticated attacker could exploit HTTP/2 flow-control by setting SETTINGS_INITIAL_WINDOW_SIZE to 0 or withholding WINDOW_UPDATE frames, causing the server to buffer complete response bodies in memory indefinitely for stalled streams. By opening many simultaneous streams requesting large resources while preventing the server from transmitting responses, an attacker could induce unbounded memory growth leading to service degradation, resource exhaustion, or denial of service. Versions v2017.01.16.00 through v2026.07.20.00 are affected.
A flaw was found in gdk-pixbuf. When parsing a specially crafted ICO file with pixel values that exceed the defined palette range, an out-of-bounds read can occur due to improper bounds checking against the actual palette size. This vulnerability causes heap bytes to be interpreted as valid palette indices and rendered as RGB pixel values in the output image, allowing an attacker to extract heap content via the generated output, such as a thumbnail.
CyberPanel through 1.9.1, fixed in commit b198460, contains an insecure direct object reference (IDOR) vulnerability in the IncBackups application's incremental-backup handlers (deleteBackup, fetchRestorePoints, and restorePoint) that allows authenticated panel users to access or manipulate other tenants' backup resources by supplying an attacker-controlled globally sequential IncJob integer ID that is never re-scoped to the authorized domain. Attackers can enumerate sequential backup IDs to read another tenant's backup metadata, irrecoverably delete another tenant's backup snapshots, or trigger unauthorized restoration of another tenant's backup job with root privileges.
CyberPanel through 1.9.1, fixed in commit b198460, contains a missing authorization vulnerability in the cancelBackupCreation handler that allows authenticated users to kill, delete, and corrupt other tenants' backups. Attackers can send crafted POST requests with arbitrary backupCancellationDomain and fileName parameters to terminate backup processes, delete backup archives, corrupt backup status files, and remove database records belonging to other tenants.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the MailInsights scheduled report configuration that allows authenticated attackers to inject arbitrary web script or HTML via the report name parameter to /Archiver/MailInsights.aspx. The injected payload is stored by ReportScheduling.btnSaveReport_Click() without output encoding and is executed in the browser of the user who created the scheduled report when they subsequently view the MailInsights page.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the default import settings configuration that allows authenticated attackers to inject arbitrary web script or HTML via the configured folders parameter to /Archiver/ImportSettingsWizard.ashx. The injected payload is stored by ImportSettingsWizard.SaveAllConfigSettings() without output encoding and is executed in the browsers of users who subsequently view the Archive Assistant default import settings.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the File Archive Assistant configuration that allows authenticated attackers to inject arbitrary web script or HTML via the excluded extensions parameter to /Archiver/FileArchiveAssistantWizard.aspx. The injected payload is stored by FileArchiveAssistantWizard.btnSave_Click() without output encoding and is executed in the browsers of users who subsequently view the File Archive Assistant settings page.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the General Settings SMTP configuration that allows authenticated attackers to inject arbitrary web script or HTML via the SMTP server address parameter to /Archiver/GeneralSettingsWizard.aspx. The injected payload is stored by GeneralSettingsWizard.SaveAllConfigSettings() without output encoding and is executed in the browsers of users who subsequently view the General Settings page.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the Call Home proxy server configuration that allows authenticated attackers to inject arbitrary web script or HTML via the proxy server address parameter to /Archiver/CallHomeSettingsWizard.aspx. The injected payload is stored by CallHomeSettingsWizard.SaveAllConfigSettings() without output encoding and is executed in the browsers of users who subsequently view the General Settings Additional Settings page.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the IMAP Server configuration that allows authenticated attackers to inject arbitrary web script or HTML via the server URL parameter to /Archiver/ImapServerWizard.aspx. The injected payload is stored by ImapServerWizard.SaveAllConfigSettings() without output encoding and is executed in the browsers of users who subsequently view the IMAP Server configuration page.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the File History Retention Policy configuration that allows authenticated attackers to inject arbitrary web script or HTML via the policy name parameter to /Archiver/FAARetentionPolicyWizard.aspx. The injected payload is stored by RetentionPolicyWizard.SaveAllConfigSettings() without output encoding and is executed in the browsers of users who subsequently view the File History Retention Policies page.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the Retention Policy configuration that allows authenticated attackers to inject arbitrary web script or HTML via the policy name parameter to /Archiver/RetentionPolicyWizard.aspx. The injected payload is stored by RetentionPolicyWizard.SaveAllConfigSettings() without output encoding and is executed in the browsers of users who subsequently view the Retention and Spam Policies page.
GFI Archiver before 15.13 contains a stored cross-site scripting vulnerability in the Classification Rules configuration that allows authenticated attackers to inject arbitrary web script or HTML via the rule name and email criteria parameters to /Archiver/CategorizationPolicyWizard.aspx. The injected payload is stored by CategorizationPolicyWizard.SaveAllConfigSettings() without output encoding and is executed in the browsers of users who subsequently view the Classification Rules page.
Improper handling of an initialization failure in AWS API MCP Server from 0.2.13 through 1.3.46 might allow an actor to bypass the user-configured security policy and execute AWS API operations that the policy was set to deny or gate. When initialization of the security policy enforcement data fails at server startup, the policy check is skipped for the lifetime of the process. IAM permissions on the configured credentials remain in effect and are unaffected.
To remediate this issue, users should upgrade to version 1.3.47.
Logto performs principal lookup without normalizing email and identifier strings, enabling principal collision and unauthorized account access via case- or Unicode-different identities.
Logto does not enforce locally configured MFA during SSO authentication, allowing users to bypass second-factor requirements and grants unauthorized access.
Logto omits validation of the SAML <Conditions> element, enabling attackers to strip time and audience restrictions and replay assertions indefinitely.
Logto bypasses OIDC nonce validation when the nonce claim is absent from the id_token, enabling replay of authentication tokens and weakening session-binding.
Logto allows unverified email-based SSO account linking, enabling an attacker to register an identity at a permissive IdP using a victim’s email and gain unauthorized access to the victim’s account.
Improper handling of insufficient permissions or privileges vulnerability in Tridium Niagara Framework on Windows, Linux, QNX, Tridium Niagara Enterprise Security on Windows, Linux, QNX allows Privilege Abuse.
This issue affects Niagara Framework: before 4.14.6, before 4.15.5; Niagara Enterprise Security: before 4.14.6, before 4.15.5.
When initializing an RSA public key from DER or PEM bytes throws an error, the EVP_PKEY* is double-freed: first in the catch block, then in the deinit. This can lead to a crash on future memory allocations. This double-free manifests when BoringSSL cannot decode the public key from the bytes provided. This vulnerability is addressed in swift-crypto version 4.5.1.
NIOSSLCertificate._subjectAlternativeNames provides access to the raw bytes for a cert's SANs. NIOSSL provides access to a buffer assumed to be backed by an ASN1_STRING, but not all SANs are backed by ASN1_STRING, so accessing the buffer for such a type can lead to out-of-bounds memory access. This vulnerability is addressed in swift-nio-ssl version 2.37.2.
Allocation of resources without limits or throttling vulnerability in BizimHesap Information Systems Industry and Trade Inc. Online Pre-Accounting Software allows Excessive Allocation.
This issue affects Online Pre-Accounting Software: through 17072026.
DOMPurify before 3.3.2 contains a mutation-XSS vulnerability when sanitized HTML is reinserted into special parsing contexts using innerHTML with wrappers like script, xmp, iframe, noembed, noframes, or noscript. Attackers can craft payloads with closing sequences that break out of the wrapper context during reparsing, reactivating dangerous markup with event handlers to execute JavaScript.
DOMPurify before 3.3.2 contains a prototype pollution vulnerability in USE_PROFILES mode that allows attackers to bypass attribute filtering by polluting Array.prototype properties. Attackers can set Array.prototype properties like onclick to true, causing DOMPurify to accept event handlers as allowlisted attributes and resulting in DOM-based XSS when sanitized markup is rendered.
DOMPurify before 3.3.2 contains a URI validation bypass vulnerability when ADD_ATTR is provided as a predicate function via EXTRA_ELEMENT_HANDLING.attributeCheck. Attackers can supply a predicate that accepts specific attribute and tag combinations to bypass URI-safe validation, allowing unsafe protocols like javascript: to survive sanitization and execute as DOM-based XSS when the link is activated.
In DOMPurify through 3.3.3, function predicates supplied via ADD_ATTR or ADD_TAGS to DOMPurify.sanitize() persist in internal state (EXTRA_ELEMENT_HANDLING) across subsequent sanitize() calls on the same instance. If a later call on the same instance provides ADD_ATTR or ADD_TAGS as an array rather than a function, the previously set function handler is neither cleared nor overwritten, so it continues to approve attacker-controlled attributes or tags. This can allow dangerous event-handler attributes or forbidden tags (bypassing FORBID_TAGS) to survive sanitization, resulting in cross-site scripting. The vendor (Cure53) considers this an edge case outside DOMPurify's threat model; the referenced advisory lists 3.4.0 as the patched version.
DOMPurify through 3.3.3 fails to sanitize DOM elements passed via IN_PLACE mode when the element originates from a different window/realm (e.g., an iframe's contentDocument). A cross-realm instanceof check in the private _isNode() function returns false for foreign-realm nodes, causing DOMPurify to stringify the element (yielding '[object HTMLDivElement]'), silently reset IN_PLACE to false, and return the unsanitized element unchanged with any XSS payloads intact. The vendor considers this an edge case outside DOMPurify's threat model and, at time of publication, no fix was planned.
DOMPurify before 3.4.0 contains a logic error in the ADD_TAGS function where short-circuit evaluation allows forbidden tags to bypass FORBID_TAGS restrictions. Attackers can craft input containing tags listed in FORBID_TAGS that are also added via ADD_TAGS function, causing them to be retained in sanitized output.
DOMPurify before 3.4.7 (affected versions <= 3.4.5) passes direct references to the module-level DEFAULT_ALLOWED_TAGS and DEFAULT_ALLOWED_ATTR sets to the uponSanitizeElement and uponSanitizeAttribute hooks via data.allowedTags / data.allowedAttributes when sanitize is called without an explicit cfg.ALLOWED_TAGS / cfg.ALLOWED_ATTR array. A hook that mutates these fields permanently widens the default allow-lists for the lifetime of the DOMPurify instance, so all subsequent default-config sanitize calls inherit the widened defaults and attacker payloads using the poisoned tag/attribute name survive sanitization. removeAllHooks(), clearConfig(), and passing a fresh cfg do not recover the state; only constructing a new DOMPurify instance does.
DOMPurify through 3.4.6 contains a cross-site scripting vulnerability in IN_PLACE mode that trusts attacker-controlled nodeName on live non-form nodes. Attackers can supply hostile live DOM objects with real script children whose observable nodeName is clobbered to appear as allowed elements, causing scripts to execute when the sanitized tree is inserted into a live document.
DOMPurify versions >=3.0.0 and before 3.4.8, when configured with SAFE_FOR_TEMPLATES together with a DOM output mode (RETURN_DOM, RETURN_DOM_FRAGMENT, or IN_PLACE), fail to strip template expressions (e.g. ${evil}, {{evil}}, <%evil%>) inside <template> element content. The final normalization/scrub pass (_scrubTemplateExpressions) uses a NodeIterator and node.normalize() that do not descend into template.content, so expressions that only form after adjacent text nodes merge survive sanitization. This bypasses SAFE_FOR_TEMPLATES and can allow a downstream template engine to evaluate attacker-supplied expressions. The string output path is not affected.
DOMPurify 3.0.0 before 3.4.9 does not reset the retained Trusted Types policy when clearConfig() is called, so a DOMPurify instance reused across trust boundaries stays bound to a previously supplied TRUSTED_TYPES_POLICY. A later caller that requests RETURN_TRUSTED_TYPE output receives a TrustedHTML object created by the old (potentially unsafe) policy rather than a clean default, which can lead to script execution at a Trusted Types sink. Passing TRUSTED_TYPES_POLICY: null on the later call also does not clear the retained policy.