Google::Auth versions before 0.09 for Perl allow server side request forgery and credential exfiltration via unvalidated URLs taken from the credentials JSON. The URLs the library requests are read from the credentials JSON, and their hosts were not checked against the universe domain before the request. For an external_account configuration, retrieve_subject_token fetched credential_source.url with headers from the same JSON, and fetch_access_token posted the subject token to token_url, then sent the STS access token it received to service_account_impersonation_url in an Authorization: Bearer header. The authorized_user, impersonated_service_account and service_account configurations posted the client secret and refresh token, the source access token, and a signed JWT assertion to their own JSON-supplied token_uri or impersonation URL. Any caller that builds credentials from a configuration it does not fully control issues those requests from the application's network position, reaching hosts the configuration names, including internal services and link-local metadata endpoints, and hands them the credentials each request carries. The service_account assertion is bound to aud, so it is not replayable against Google. Version 0.06 added a _validate_url host check to the external_account class, keyed on a universe_domain read from the same credentials JSON. Version 0.07 gated a JSON-supplied universe domain behind GOOGLE_EXTERNAL_ACCOUNT_ALLOW_CUSTOM_UNIVERSES=1, deriving the pin flag from arguments that an earlier BUILDARGS pass had already merged on the make_creds path. Version 0.08 passed the pin decision through as an explicit constructor argument and moved _validate_url to Google::Auth::Credentials, adding the call to UserRefreshCredentials and ImpersonatedServiceAccountCredentials, and 0.09 added it to ServiceAccountCredentials.

Flowise is a drag & drop user interface to build a customized large language model flow. Prior to 3.1.3, the PUT /api/v1/executions/:id endpoint in packages/server/src/routes/executions/index.ts lacks the checkAnyPermission() middleware that protects other execution endpoints. Any authenticated user, regardless of assigned permissions, can modify execution state, data, and metadata of any execution in their workspace, enabling privilege escalation and manipulation of workflow execution results. This issue is fixed in 3.1.3.

Flowise is a drag & drop user interface to build a customized large language model flow. Prior to 3.1.3, Flowise validatePythonCodeForDataFrame in packages/components/src/pythonCodeValidator.ts can be bypassed with Unicode homoglyph identifiers, allowing arbitrary Python execution inside Pyodide and full OS command execution on the Flowise host via Pyodide js module interop. The validator gates pyodide.runPythonAsync in packages/components/nodes/agents/CSVAgent/CSVAgent.ts and packages/components/nodes/agents/AirtableAgent/AirtableAgent.ts with an ASCII word-boundary blacklist. JavaScript regex word boundaries are ASCII-only, while Python 3 NFKC-normalizes identifiers at parse time, so homoglyph forms such as __cl𝐚ss__, __subcl𝐚sses__, __b𝐚se__, and __b𝐮iltins__ bypass the blacklist and are parsed as their ASCII equivalents. This issue is fixed in version 3.1.3.

Flowise is a drag-and-drop user interface for building customized large language model (LLM) flows. Prior to version 3.1.3, several custom-tool components — AgentAsTool, ChatflowTool, and ExecuteFlow — ran code in the in-process  vm2  sandbox. To build that code, they inserted a user-controlled  baseURL  value straight into the JavaScript source, for example  const url = "${baseURL}/..."; . The only check on  baseURL  was  isValidURL , but a valid-looking URL can still contain characters that break out of a code string. An authenticated user could craft a  baseURL  that passed this check, closed the surrounding string, and injected their own JavaScript into the sandboxed script (code injection, CWE-94). The  vm2  sandbox runs in the same Node.js process as Flowise and exposes risky dependencies. As a result, the injected code could escape the sandbox and run arbitrary code on the Flowise server as the Flowise process user. Exploitation only requires an authenticated session. The issue is fixed in version 3.1.3, which passes the URL to the sandbox as data instead of inserting it into code and adds stricter URL validation.

The fix released in jackson-core 2.18.6 and 2.21.1 for CVE-2026-18401 (GHSA-72hv-8253-57qq, number length constraint bypass in the non-blocking parser) is incomplete. This record covers the remaining bypass. The earlier fix wired validateIntegerLength() into a new _setIntLength() helper and invoked it wherever the integer portion of a number is decided: a terminator byte arrives, a '.' or 'e'/'E' is seen, or input ends inside a fully buffered value. It was not invoked on the attacker-relevant path where the parser runs out of input while still inside the MINOR_NUMBER_INTEGER_DIGITS minor state and returns NOT_AVAILABLE to the caller. As a result, an attacker who streams JSON to a non-blocking parser in many small chunks, without ever sending a terminator byte, keeps the parser inside MINOR_NUMBER_INTEGER_DIGITS indefinitely. _textBuffer.expandCurrentSegment() grows the accumulator on every chunk while validateIntegerLength() is never called. The accumulator is bounded only by maxStringLength (20 MiB by default) rather than by maxNumberLength (1000 by default), an amplification of roughly 20,000x over the documented limit. Because Java char values occupy two bytes, a single connection can be driven to approximately 40 MiB of heap before the validator finally fires when the value completes. The equivalent fraction-path code is correct: _finishFloatFraction() calls _setFractLength() before its NOT_AVAILABLE return. The missing call affects the integer-digit paths in _startPositiveNumber(), _startNegativeNumber() and _finishNumberIntegralPart() in NonBlockingUtf8JsonParserBase. Impact: reactive frameworks such as Spring WebFlux/Reactor, Quarkus, Helidon and Vert.x feed inbound HTTP or gRPC bytes to the async parser as they arrive, which is precisely the chunked-feed shape required. Operators who set StreamReadConstraints.maxNumberLength expecting it to cap memory per number value do not get that guarantee; memory accumulates per concurrent connection and attacker-controlled concurrency can exhaust the JVM heap. The synchronous parsers (UTF8StreamJsonParser, ReaderBasedJsonParser) and the async parser operating on complete input are not affected. Exploitation requires only the ability to stream data to a parsing endpoint; no privileges or user interaction are needed. This issue affects com.fasterxml.jackson.core:jackson-core from version 2.15.0 through 2.18.7, and from 2.19.0 through 2.21.3, and tools.jackson.core:jackson-core from 3.0.0 through 3.1.3. Versions prior to 2.15.0 are not affected, because StreamReadConstraints -- which defines the maxNumberLength setting -- was first introduced in jackson-core 2.15.0, so no such constraint exists to be bypassed in earlier releases. Note that GHSA-r7wm-3cxj-wff9 states the affected 2.x range without a lower bound. The 2.22.x and 3.2.x release lines are not affected: those branches were created after the fix commit landed on 2026-05-21 and therefore contain it from their initial releases (2.22.0, tagged 2026-06-03, and 3.2.0, tagged 2026-06-08).

Cross-Site Request Forgery vulnerability in Erlang Ecosystem Foundation oidcc_plug (Oidcc.Plug.AuthorizationCallback module) allows an attacker to make a victim's browser complete an authorization flow the victim never initiated. This vulnerability is associated with program file lib/oidcc/plug/authorization_callback.ex and program routine Oidcc.Plug.AuthorizationCallback.call/2. A callback request that carries no Oidcc.Plug.Authorize session is processed with every security check disabled rather than being rejected. call/2 substitutes permissive defaults for the absent session, and each downstream check treats its value as nothing to compare and returns :ok, so the nonce, state, PKCE, peer IP and user agent checks are all skipped. A separate clause of check_state/2 also accepts a state-less request when a verifier is present. An attacker obtains an authorization code for their own provider account, then induces the victim to visit the callback endpoint with that code and no state parameter. The application signs the victim in as the attacker, so the victim's subsequent actions occur in the attacker's account where the attacker can read them. Applications reusing one callback for both signing in and linking a provider account are further exposed to account takeover, the attacker's account becoming linked to the victim's. The permissive fallback serves no conforming flow. Third-party-initiated login reaches a relying party at a separate login initiation endpoint and causes it to send a fresh authentication request, and this library implements no such endpoint. Oidcc.Plug.Authorize always sends a state parameter, which an authorization server must echo, so no legitimate callback lacks one. This issue affects oidcc_plug: from 0.2.0-beta.1 before 0.5.0.

In the Linux kernel, the following vulnerability has been resolved: KVM: x86: Check for invalid/obsolete root *after* making MMU pages available Check for a "stale" page fault, i.e. for an invalid and/or obsolete root, after making MMU pages available for the shadow MMU. If reclaiming shadow pages zaps an in-use root, i.e. marks it invalid, then KVM will attempt to map memory into an invalid root. On its own, populating an invalid root is "fine", but because child shadow pages inherit their parent's role, any children created during the map/fetch will be created as invalid pages, thus violating KVM's invariant that invalid pages are never on the list of active MMU pages. Note, the underlying flaw has existed since KVM first started tracking invalid roots in 2008 (commit 2e53d63acba7, "KVM: MMU: ignore zapped root pagetables"), but the true badness only came along in 2020 (Linux 5.9) with the invariant that invalid shadow pages can't be on the list of active pages. Note #2, inheriting role.invalid when creating child shadow pages is also far from ideal; that flaw will be addressed separately.

The Clearfy Cache WordPress plugin before 2.4.3 does not perform a capability check in one of its admin-page dispatch paths, allowing any authenticated user such as a Subscriber to render admin-only settings pages and disclose their contents, including administrative nonces, while the canonical page URL correctly restricts access.

The miniOrange 2FA WordPress plugin before 6.2.7 does not restrict who can trigger its second-factor configuration OTP send, nor bind the OTP recipient to the enrolling user's own address, allowing a low-privileged user to send one-time-passcode emails to arbitrary recipients and to exhaust the site's metered OTP allowance, preventing legitimate users from receiving their second-factor codes.

The wpForo Forum WordPress plugin before 3.1.3 does not restrict which profile fields a member may set when editing their own account, allowing users with a subscriber-level account to write administrator-controlled account-state and reputation fields on their own profile, including self-activating a pending or banned account and forging their forum reputation score.

A flaw was found in the backchannel logout endpoint of the keycloak-services component, which is part of the Red Hat Build of Keycloak. This component handles authentication and session management for applications. The issue occurs when an OIDC identity provider is configured to skip signature validation. In this specific setup, the system incorrectly accepts logout requests that have no cryptographic signature. An attacker who knows certain technical details about a user's session can use this flaw to force that user to be logged out, potentially disrupting their work.

Ouroboros is a local-first runtime for AI coding agents that records their actions and applies user-defined policies to constrain behavior. Versions prior to 0.42.1 have an incomplete denylist. Several execution-routing keys of the same RCE class were omitted, so a malicious cloned repo can still reach arbitrary command execution by shipping a .env (auto-loaded at import, with no review step). The CVE-2026-47211 fix added _UNTRUSTED_ENV_DENYLIST to stop an untrusted project-directory .env from redirecting execution, but it did not account for all keys. The backend config-home and MCP/plugin roots bypass the approval gate by pointing the nested agent, MCP servers, and plugin roster at attacker config. Other variables re-enable blocked local transports, replace sub-agent prompts, switch backends, and lower tool approval classes, further weakening the approval gate. This issue has been fixed in version 0.42.1.

Ouroboros is a local-first runtime for AI coding agents that records their actions and applies user-defined policies to constrain behavior. In versions prior to 0.39.0, if a user clones a malicious repository and runs Ouroboros commands within that directory, it can lead to arbitrary code execution and potential system takeover. The vulnerability stems from Ouroboros loading the .env file from the current working directory. Execution-affecting environment variables such as OUROBOROS_CLI_PATH, OPENCODE_CLI_PATH, and other backend selectors are accepted directly from this local .env. An attacker can include a malicious script in the repository and point the CLI path variable to it (e.g., OUROBOROS_CLI_PATH=./malicious_script.sh). When the user executes a command like ouroboros init or any command that instantiates the adapter, the malicious script is executed instead of the intended CLI. This issue has been fixed in version 0.39.0.

ArcadeDB before 26.7.3 binds the real LocalDatabase object into JavaScript trigger contexts with HostAccess.ALL, allowing schema-admins to call getSecurity().createUser() without permission checks. Attackers with UPDATE_SCHEMA permission can create triggers that execute JavaScript to create server-wide admin users, escalating privileges beyond their authorization level.

Keycloak provides authorization services that allow administrators to restrict access to resources based on time policies (for example, only allowing access during business hours). A flaw was discovered where a user can include a fake time value in their authorization request that overrides the actual server time. This allows the user to bypass these time-based restrictions and access protected resources at unauthorized times.

The SMS Alert WordPress plugin before 3.9.8 does not bind its "mobile verified" session flag to the phone number that was actually verified: after an attacker verifies an OTP sent to their own phone, the signup/login handler reads a fresh, attacker-supplied phone number to select the account and logs them in. An unauthenticated attacker can therefore log in as any user, including an administrator, who has a billing phone on file.

better-auth SCIM versions from 1.5.0 before 1.7.0-beta.4 fail to bind non-organization SCIM providers to their creator by default, allowing authenticated users to manage other users' providers. Attackers can regenerate SCIM bearer tokens, invalidate legitimate tokens, and authenticate to SCIM API routes with the attacker-controlled token.

@better-auth/stripe versions >= 1.4.11 and < 1.6.21, and >= 1.7.0-beta.0 and < 1.7.0-beta.10, contain an authorization bypass in organization subscription actions. The middleware validates the organization ID taken from the request query string against the authorizeReference callback, but the handler reads the organization ID only from the request body and falls back to the caller's active organization from their session. When these differ, an authenticated member of multiple organizations can perform subscription actions (cancel, change plan, restore, billing portal access) against an organization they belong to but should not manage, and can access another organization's billing details including payment methods, invoices, and subscription state.

OpenRemote (org.openremote:openremote) versions <= 1.26.2 contain an insecure direct object reference vulnerability in the setAssetLinks endpoint of AlarmResourceImpl. The realm access check validates only a single realm obtained via realms.stream().findFirst() on a HashSet of realms from the request, rather than all realms. Because HashSet iteration order is non-deterministic, an authenticated attacker who includes alarm-asset links from both their own realm and a victim realm can, with roughly 50% probability per request (retryable), persist cross-tenant links and disclose victim asset names (returned via @Formula fields) through GET requests on the attacker's own alarm. Fixed in 1.27.0.

The Pronamic Pay plugin for WordPress is vulnerable to Privilege Escalation in all versions up to, and including, 10.1.0 This is due to the `maybe_update_user_role()` function passing an attacker-controlled Gravity Forms field value (`$lead[$feed->user_role_field_id]`) directly into `WP_User::set_role()` without any allowlist validation, capability comparison, or permission check to constrain which roles can be assigned. This makes it possible for authenticated attackers, with Subscriber-level access and above, to escalate their own WordPress account to Administrator by tampering with the role field value in a form submission. Exploitation requires that an administrator has already configured a Pronamic Pay payment feed in Gravity Forms with the **Update User Role** option enabled and mapped to a form field; once that configuration is in place, no further preconditions exist to prevent an authenticated attacker from exploiting this vulnerability.

The Chat On Desk Order Notifications WordPress plugin before 1.0.9 does not verify that the one-time password has been validated before processing a password-reset request, allowing unauthenticated attackers to reset the password of arbitrary users, including administrators, and take over their accounts when SMS one-time-password password reset is enabled.

The Fluent Support WordPress plugin before 2.3.1 does not perform a per-ticket access check before reassigning a ticket's customer, allowing a restricted support agent to change the assigned customer of any ticket in the system, including tickets outside their granted scope.