A flaw was found in the Keycloak Admin REST API, which is used to manage security realms and clients. The issue occurs when the system processes requests for rotated client secrets that are stored in a secure vault. Due to improper boundary enforcement, a delegated administrator with view-only permissions can retrieve the actual resolved secret instead of the vault placeholder, leading to the exposure of sensitive credentials.

Parse Server versions >= 9.0.0 before 9.10.0-alpha.5 and >= 8.2.2 before 8.6.86 return GraphQL validation error messages that name required custom input fields even when public introspection is disabled (graphQLPublicIntrospection: false, the default). A client holding only the public application id — with no user session, master key, or maintenance key — can trigger validation errors to learn the names of required (non-null) custom fields on classes it already references by name, partially defeating the schema-hiding intent of disabling public introspection. No stored data, credentials, optional field names, unreferenced class names, or Cloud Code function names are exposed.

Parse Server versions >= 9.0.0 before 9.10.0-alpha.6 and >= 8.2.2 before 8.6.87 disclose Pointer and Relation target class names through GraphQL validation and input-coercion error messages when public schema introspection is disabled (graphQLPublicIntrospection: false, the default). Because these errors are produced before authentication, authorization, or any resolver runs, an unauthenticated client possessing only the public application ID can trigger errors on Pointer or Relation fields to reconstruct hidden schema class names, partially defeating the schema-hiding protection. Only schema metadata (class names) is exposed; no object data, credentials, or user records are disclosed.

Apereo CAS Client accepts any CA-trusted certificate for any hostname, provided the URL the client is calling matches the configured allowlist or regex. An attacker with a MITM position (DNS poisoning, rogue Wi-Fi, malicious proxy, etc.) can provide any CA-signed certificate for a hostname that matches the configured allowlist or regex. This can lead to intercepting the CAS exchange, capturing the Ticket-Granting Ticket (TGT), and subsequently obtaining Service Tickets on behalf of the victim.  Because maintainers contact attempts were unsuccessful, vulnerabilities have only been confirmed in version 4.1.0 (Java Apereo CAS Client) and 3.6.4 (Jasig CAS Client) but may also affect other versions.

Multiple security vulnerabilities in Snowflake libsnowflakeclient versions prior to 2.9.2 could allow remote code execution and credential exfiltration. A stack-based buffer overflow in the file download path could allow remote code execution on a victim host. An attacker could exploit this by uploading a file with a crafted encryption metadata field to a shared internal stage that a victim process later downloads, and impact would be limited to deployments where principals with different privilege levels share the same internal stage. A related out-of-bounds write in the same download path could allow memory corruption with attacker-controlled write primitives. An attacker may exploit this through a crafted initialization vector metadata field on a shared stage, and impact would be limited by the same stage-write precondition. Improper validation of connection parameters could allow an attacker-controlled input to redirect outbound authentication requests — including credentials and tokens — to an attacker-controlled endpoint. Impact is limited to embedding deployments where a lower-privileged principal can influence connection configuration while higher-privileged service credentials are in use. The fix is available in Snowflake libsnowflakeclient version 2.9.2. The Snowflake PHP PDO Driver and Snowflake ODBC Driver embed the affected library; fixes are available in versions 4.1.0 and 3.19.0 respectively. Users must manually upgrade.

A vulnerability in zenml-io/zenml versions 0.57.0 through 0.94.2 allows an attacker to bypass rate-limiting on the `POST /api/v1/login` and self password-change endpoints by rotating the `X-Forwarded-For` header. The rate limiter keys requests by `request.client.host`, which is derived from the `X-Forwarded-For` header when Uvicorn is launched with `--proxy-headers --forwarded-allow-ips *`. This configuration allows clients to control the value of `request.client.host`, effectively bypassing rate-limiting protections. This vulnerability leaves the affected endpoints open to unthrottled credential guessing attacks.

Shopware is an open commerce platform. Versions 6.7.3.0 through 6.7.10.0 have an open redirect in Shopware's public SSO entry point at `GET /api/oauth/sso/auth`. When the endpoint is reached without the expected SSO session state, the application falls back to the request's `Referer` header and uses that value as the redirect destination. In the validated behavior, the server does not restrict that fallback target to same-origin URLs, does not require a relative path, and does not reject dangerous schemes such as `javascript:`. As a result, an unauthenticated request can turn this endpoint into a reusable redirect primitive whose destination is fully controlled by attacker-supplied request metadata. The security problem is not limited to a harmless navigation mismatch. The endpoint sits under `/api/oauth/`, which gives the redirect a trustworthy application-controlled origin and makes it suitable for phishing chains, branded redirect abuse, and cases where client software automatically follows redirects issued by a trusted host. The attached evidence also shows that the response is not only an HTTP `302` with a user-controlled `Location` header. The HTML body contains a matching meta refresh tag and redirect link built from the same attacker-controlled value. In the validated proof, the endpoint redirects to `https://attacker.example/poc` when that URL is supplied through `Referer`, and it also reflects `javascript:alert(1)` into `Location` and the HTML redirect body without any scheme filtering. This report therefore stays conservative and claims an open redirect with arbitrary redirect targets, while noting that the lack of scheme restrictions makes the behavior materially worse than a same-scheme external redirect. Version 6.7.10.1 fixes the issue.

A security issue was discovered in the Kubernetes Java client library where a compromised pod may be able to create new files in arbitrary locations on the client machine executing copy operations via non-tar copyDirectoryFromPod when enableTarCompressing is false.

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.

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.

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.

Joomla Extension - regularlabs.com - IP spoofing vulnerability in GeoIP extension - GeoIP lookups trusted spoofable forwarded client-IP headers, this could cause GeoIP-rule bypass.

Joomla Extension - regularlabs.com - IP spoofing vulnerability in IP login extension - IP Login trusted forwarded client-IP headers without requiring a configured trusted proxy. Attackers could spoof the IP used for automatic login and potentially impersonate mapped accounts.

Joomla Extension - regularlabs.com - Client IP spoofing vulnerability in Regular Labs conditions manager - IP and GeoIP conditions trusted spoofable forwarded headers, allowing remote clients to bypass location-based rules.

A flaw was found in librest. The PKCE implementation for OAuth authorization uses the GRand function from the GLib API, a cryptographically insecure pseudo-random number generator. Because the generated "code verifier" lacks sufficient cryptographic entropy, a malicious actor can reverse-engineer the pseudo-random number generator (PRNG) seed to predict or reconstruct the code verifier string, allowing an attacker to bypass PKCE protections and successfully impersonate the client during the OAuth 2.0 authorization flow.

When PROXY protocol v2 is used on the Unix domain socket path, roles derived from X.509 client certificates may not be validated against the configured tlsCATrusts allow-list. This can result in unintended role assignments following MONGODB-X509 authentication. Affected scenarios require local access to the proxy Unix domain socket and a valid X.509 certificate issued by a trusted certificate authority.

mongodb

Improper handling of DBPointer objects during BSON serialization in MongoDB's server-side JavaScript engine can result in internal process memory contents being included in data returned to the client. This constitutes an unintended information disclosure affecting deployments that use server-side JavaScript.

mongodb

An authenticated user with low privileges may be able to perform unauthorized reads and writes on data protected by role-based query-level access controls, due to insufficient validation of certain client-supplied command parameters. The issue affects find, update, delete, and aggregate commands in non-apiStrict configurations.

mongodb

An issue in the server’s Atlas Search integration allows an authenticated user to bypass per-user access controls. In sharded topologies, the $search and $searchMeta aggregation stages use internal routing that is normally populated only by the trusted router during sharded search planning. Due to insufficient input validation, an authenticated client can supply these fields directly.

mongodb

An authentication bypass vulnerability in Check Point Security Management and Multi-Domain Security Management allows an unauthenticated remote attacker to execute administrative commands on the Management Server. Successful exploitation may also allow command execution on managed Security Gateways. Exploitation requires network access to the Management Server without firewall protection or a configuration that does not restrict Trusted Clients.