Unauthenticated visitors can overwrite a machine-translation configuration value in the Infility Global WordPress plugin before 2.16.25, because an AJAX action exposed to logged-out users performs no authorization or nonce check (CWE-862). The impact is integrity-only and low: no confidentiality or availability effect is described, and the affected setting belongs to a companion Infility Global plugin. Severity is low (CVSS 3.7; assessed CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N), the vendor has released 2.16.25, a public PoC reference exists but is not verified as working exploit code, and the vuln.today decision label is monitor.
Predictable session identifiers in the Linksys RE6250 Web Management Interface (firmware 1.0.04.001) can expose session-scoped information because session_id values are generated with insufficient randomness and checked with strcmp. A remote, unauthenticated attacker could exploit this only by successfully predicting a valid token, a high-complexity condition; publicly available exploit code exists, and the impact is limited to low confidentiality with no integrity or availability effect (CVSS 4.0 base 6.3; assessed CVSS 3.1 C:L/I:N/A:N).
Broken access control in the MasterStudy LMS WordPress plugin (versions 3.7.13 up to but not including 3.7.55) lets an authenticated user who holds the Instructor role rename any account on the site - first name, last name and public display name - including administrator accounts, opening the door to impersonation and phishing. The plugin's student-enrolment feature does not check that the targeted user has any relationship to the requesting instructor, so the only barrier is obtaining Instructor-level access; there is no confidentiality or availability impact and no ability to change credentials or privileges. Severity is low (assessed CVSS 2.7, PR:L), there is no CISA KEV entry or other signal of active exploitation, and a public PoC reference exists but is not verified as working exploit code; priority is monitor.
The MB Custom Post Types & Custom Taxonomies plugin before 2.12.3 allows any authenticated WordPress user with the Contributor role to use its export feature without a capability check, disclosing admin-only custom post type and taxonomy configuration plus titles, dates, statuses and some stored content from other users' non-public posts and pages. Impact is read-only and low severity, and exploitation requires a valid low-privilege account on an affected site. A public PoC reference exists, but it is not verified as working exploit code.
Password handling in patientregistration.php of subhajitkhan's online-clinic-management-system (through commit e9ee77a8827a1446220fa07ee693dc4d9a29a578) fails to encrypt the pat_pswd and doc_pswd credential values, exposing sensitive data (CWE-311). The vendor's CVSS 4.0 score of 5.3 reflects a network-reachable but low-privilege-authenticated precondition (PR:L, no user interaction); our independent assessment rates it CVSS 3.1 5.3 with only low confidentiality impact and no integrity or availability effect. No CISA KEV entry exists and no public exploit has been verified - the referenced GitHub issue #9 is a public PoC reference only and is not confirmed working exploit code - so the practical risk is constrained by the project's small deployment footprint and the low impact, keeping this at routine, non-emergency priority.
Online Clinic Management System (subhajitkhan/online-clinic-management-system) fails to enforce its appointment-booking workflow correctly, allowing a low-privilege authenticated user to manipulate requests to the patappointtime.php Booking Handler and alter booking state. The project is a niche self-hosted PHP application, and the CVSS 4.0 base score of 5.3 reflects low integrity and availability impact confined to the booking workflow, with no confidentiality exposure. A public PoC reference exists in the form of an upstream GitHub issue; it is not verified as working exploit code, and there is no confirmed active exploitation. The vendor was notified early via an issue report but has not responded, and the fix is identified as commit e9ee77a8827a1446220fa07ee693dc4d9a29a578.
Improper input validation in the Appointment Booking component of subhajitkhan's Online Clinic Management System (source tree up to commit e9ee77a) lets a logged-in low-privilege user tamper with the day, time and appdate parameters handled by patappointtime.php. The result is low-integrity and low-availability impact on booking data with no confidentiality exposure, rated 5.1 under CVSS 4.0; exploitation requires an authenticated booking-capable account and some user interaction. The project maintainer was notified through an issue report but has not responded, and the referenced public PoC is an unverified issue-report link rather than confirmed working exploit code, so no public working exploit has been established at time of analysis.
Cross-site scripting in the 'Query' argument handled by patdocmsg.php in subhajitkhan's online-clinic-management-system (all code through commit e9ee77a8827a1446220fa07ee693dc4d9a29a578) allows an authenticated low-privilege user to inject script that executes in another user's browser when that second user views the affected page. The vendor has been notified through a public issue report but has not responded, and no fixed version is available; a public PoC reference exists (GitHub issue #8), but it is not verified as working exploit code and there is no confirmed active exploitation (CISA KEV). Priority is low - the project is an obscure individual GitHub repository, authentication and victim interaction are both required, and the assessed impact is confined to low integrity with no confidentiality or availability loss.
Secrets used by the HCL AppScan Azure Boards Plugin can be disclosed in plaintext during Azure DevOps pipeline execution when pipelines are configured to emit task outputs, debug or verbose logs, environment-variable dumps, or build-agent scripts. The issue affects pipelines that use the plugin and requires an authenticated actor with sufficient pipeline privileges to configure or run the pipeline and/or read its output. No public exploit was identified at time of analysis, and available signals support a low-priority monitor decision because exposure depends on specific non-default pipeline configurations.
Authentication bypass in the Drupal contributed module Stop Administrator Login 1.0 through 1.6 lets an authenticated high-privilege user reach a protected login path through an alternate route that the module does not cover, yielding partial confidentiality and integrity impact. The vendor has released a patch referenced by Drupal SA-CONTRIB-2026-183, EPSS is 0.18% (7th percentile), and no public exploit has been identified at time of analysis. Consistent with the CVSS 3.8 score and monitor priority, real-world risk is low and exploitation requires existing high-privilege access.
The Drupal Mailer Plus Log contributed module writes sensitive information into its log files, exposing that data to anyone who can read the logs, up to and including version 1.2.7. Exploitation requires the module to be deployed on a Drupal site and an attacker who already holds high-privilege authenticated credentials able to reach the module's logging functionality, so this is a post-compromise information-disclosure issue rather than an anonymous-entry vector; the exact sensitive values logged and the operation that writes them are not described in the available sources. All signals converge on low real-world priority: EPSS is 0.15% (4th percentile), CISA SSVC records no known exploitation and no automatable path, no public exploit code is identified, and a vendor patch is available.
Drupal sites running the SAML SSO - Service Provider contrib module at versions up to and including 3.2.0 store sensitive information insecurely, allowing a highly privileged authenticated user to retrieve embedded sensitive data. The flaw is network-reachable but requires high privileges (PR:H) on the Drupal site, and the vendor has released a patched version; no confirmed active exploitation or working exploit code was observed, and EPSS is 0.15% (4th percentile). Exploitation preconditions narrow the realistic exposure to trusted administrative accounts or already-compromised privileged sessions on sites where the module is installed and enabled.
Memory exhaustion in Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 occurs when thumbnail generation falls back to PHP GD on non-Windows servers, where a crafted image with oversized dimensions is processed without the memory check that previously applied only on Windows. The flaw is availability-only, requires the GD fallback to be in effect and a victim to trigger thumbnail rendering (for example by opening a crafted message or attachment), and carries a low CVSS score of 3.7. No public exploit code or confirmed active exploitation was identified at time of analysis.
Denial of service in mod_qos can crash an Apache worker process when the module inspects HTTP response bodies and processes a specifically crafted server response. The flaw is an out-of-bounds read (CWE-125) reachable by unauthenticated remote attackers, but exploitation requires the non-default response-body inspection feature and a deployment where response content is attacker-influenced, so practical risk is low. No public exploit or CISA KEV listing was identified in the provided data.
Denial of service in mod_qos can crash an individual Apache worker process when the module's non-default JSON body filtering is enabled, triggered by a crafted percent-encoded JSON request that exploits improper string termination (CWE-170) and causes an out-of-bounds read. The CVSS 3.1 vector is AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:L (score 3.7), so the issue is remotely reachable without authentication but has high attack complexity and only low availability impact. No public exploit code or CISA KEV active-exploitation status is identified in the provided intelligence.
Roundcube Webmail before 1.6.20 and 1.7.x before 1.7.5 contains a CSRF validation bypass in the markasjunk plugin's mark-as-junk action, allowing a forged request to mark a logged-in victim's messages as junk without a valid CSRF token. The attacker must induce the victim to trigger the cross-site request while an authenticated webmail session is active, and the markasjunk plugin must be enabled. Impact is limited to a low-integrity mailbox state change; no public exploit identified at time of analysis, and the assessed CVSS 3.1 base score is 3.7.
Missing authorization in the Islandora digital repository framework for Drupal (versions 0.0.0 through 2.19.0) lets unauthenticated, network-positioned attackers reach protected resources by requesting their paths directly - a forceful browsing bypass of access control. Impact is limited to low-severity confidentiality disclosure with no integrity or availability effect, and the trigger specifics (which endpoints or object states bypass the check) are not described in the available sources. No public exploit code and no active exploitation are confirmed; EPSS is 0.14% (3rd percentile), CISA SSVC records Exploitation: none and Automatable: no, and the vendor has released a fix, so priority is monitoring rather than emergency response.
Signature spoofing in the Drupal Monobank payment API module (versions 0.0.0 through the 1.0.3 range boundary) allows a remote, unauthenticated actor to submit a payment notification/callback whose cryptographic signature the module fails to verify properly, so a forged payment confirmation can be accepted as genuine. Impact is limited to integrity (CVSS 3.7, C:N/I:L/A:N) with no confidentiality or availability effect, and only sites that have the contrib module installed and active as a payment integration are exposed. Exploitation status is low-risk: no CISA KEV listing, EPSS 0.10% (1st percentile), a public PoC reference exists but is not verified as working exploit code, the vendor has released a fixed release, and the independent assessment rates all decision signals as a low real-world priority (monitor).
libslirp user-mode networking can disclose the host's real IP address to a guest when an ICMP echo request is handled through the fallback path and the reply length is not updated to match the actual data received. The flaw affects libslirp as shipped in Red Hat Enterprise Linux 8, 9, and 10 and Red Hat OpenShift Container Platform 4, with exposure limited to guests using slirp-style user-mode networking rather than bridged or tap networking. Rated CVSS 3.5 with a monitor priority, exploitation requires a guest-level user able to send ICMP echo requests plus a host-side condition the attacker does not control, and no public exploit has been identified at time of analysis.
Mermaid Diagram Field 1.0.4 to 1.0.9, a contributed Drupal module, contains an access-control gap (CWE-863) that lets a user who already holds a valid low-privileged account reach Mermaid diagram field content they are not permitted to view by requesting it directly (forceful browsing). The result is limited confidentiality exposure of diagram content only - integrity and availability are unaffected - and the CVSS vector (AV:N/AC:L/PR:L/UI:R, CVSS 3.5) models both an authenticated attacker and a user-interaction component. No public exploit code has been identified at time of analysis, EPSS is 0.18% (7th percentile), and the case is not in CISA KEV; the vuln.today decision label is monitor, consistent with a low real-world priority.
Script injection in the Drupal Webform contributed module (versions 0.0.0 through 6.2.12 and 6.3.0 through 6.3.1) lets an actor who already holds high Webform privileges place malicious script that executes in another user's browser when that user interacts with the affected page. The injection point is a webform element or configuration setting that the available sources do not identify, and successful exploitation requires both the elevated privilege and victim interaction, keeping confidentiality and integrity impact low and scope unchanged. This is a low-severity issue: EPSS is 0.17% (6th percentile), it is not in CISA KEV, no public exploit code is confirmed, and the vendor has published a patch; the current priority decision is monitor.
Template injection (CWE-1336) in the Drupal AI CKEditor module affects versions 0.0.0 through 1.4.3, allowing an authenticated Drupal user who can use the module's features to inject unneutralized template syntax that is later evaluated by a template engine, producing limited code injection. Successful exploitation requires the victim to interact with crafted content (CVSS PR:L/UI:R), and impact is confined to low confidentiality and integrity effects within the unchanged component scope. No public exploit code or active exploitation has been reported, and availability of an updated release (1.4.3) keeps priority at monitor level.
Cross-site scripting in the Drupal contributed module SAML SSO - Service Provider affects versions 0.0.0 through 3.2.0 and allows an attacker holding a high-privilege account on the Drupal site to get script executed in another user's browser when that user loads the page carrying the injected content. Scoring is modest (CVSS 3.1 3.5, AV:N/AC:L/PR:H/UI:R/S:U/C:L/I:L/A:N) because the attacker must already be elevated and a separate victim must interact, with only low confidentiality and low integrity impact and no availability effect. No public exploit code or CISA KEV entry is associated with the issue, EPSS sits at 0.16% (5th percentile), and the priority label is monitor; a vendor patch is available, with the Drupal advisory sa-contrib-2026-146 as the authoritative reference.
Incorrect authorization in Drupal Webform allows forceful browsing; all signals point to low real-world priority. Affected releases are the supplied 0.0.0-6.2.12 and 6.3.0-6.3.1 ranges, with the independent assessment identifying 6.2.12 and 6.3.1+ as not affected, and successful exploitation is limited to low confidentiality and integrity impact. No public exploit was identified at time of analysis, EPSS is 0.18% (7th percentile), and the vendor patch is available; the supplied vendor vector requires high privileges and high attack complexity, while the independent reassessment lowers those requirements to low privileges and low complexity.
Cross-site scripting (CWE-79) in the Drupal contrib module SAML SSO - Service Provider, affecting every release from 0.0.0 through 3.2.0, lets an attacker-supplied payload be rendered into a web page without proper neutralization, with low confidentiality and integrity impact. The vendor CVSS vector (AV:N/AC:H/PR:H/UI:N) requires an authenticated high-privilege account and high attack complexity, which keeps the practical exposure narrow; EPSS sits at 0.16% (5th percentile) and no public exploit identified at time of analysis, so this is a low-severity issue rather than a patching emergency. A vendor fix is available, and sites running the module should schedule the update through the normal contrib maintenance cycle.
Server-side request forgery in the Drupal SAML SSO - Service Provider module affects versions 0.0.0 through 3.2.0 and allows a highly privileged account that can influence the module's SAML/IdP configuration to make the Drupal server send requests to attacker-influenced destinations. Impact is limited to low confidentiality and low integrity with no availability effect, and exploitation also requires high attack complexity. EPSS is 0.15% (4th percentile), no public exploit identified at time of analysis, and the priority is monitor.
HPE networking products can disclose limited confidential data to a local, already-authenticated low-privileged user who also completes user interaction, with minor availability impact and no integrity effect. The exact affected product, version range, and triggering feature are not described in the available sources; the HPE security bulletin is the primary reference. No public exploit code or CISA KEV confirmation is available, and the issue is rated low priority for monitoring.
Roundcube Webmail versions before 1.6.20 and 1.7.x before 1.7.5 can allow an authenticated user to reach an action that an administrator intentionally placed in the non-default disabled_actions list by invoking that action through a deprecated alias. The pre-fix authorization check compared only the requested action name and did not resolve aliases first, so the restriction can be bypassed; the assessed impact is integrity-only, with no confidentiality or availability effect. Exploitation requires the disabled_actions setting to be explicitly configured, valid credentials, and knowledge of the deprecated alias; no public exploit identified at time of analysis and no CISA KEV entry is present in the provided data.
Supabase Realtime versions up to and including 2.102.2 accept an expired JWT on its REST broadcast endpoint when the token's exp claim is encoded as a JSON string rather than a JSON number, allowing a token holder to keep publishing broadcast messages to subscribers after the stated expiry. Only the REST /api/broadcast authorization path is affected - the WebSocket channel-join path already rejects the same token because it additionally type-checks exp as an integer - and signature verification against the tenant secret, tenant-secret selection and any configured role validator still apply, so the real impact is limited to message injection whose consequences are application specific. The issue is rated low severity (CVSS 3.1, PR:L/AC:H), no public exploit code was identified at the time of analysis, and the fix ships in version 2.103.1.
Capture-replay of a valid SAML assertion can bypass authentication in the Drupal SAML SSO - Service Provider module for versions 0.0.0 through 3.2.0, allowing an attacker who has obtained a legitimate SAML response to replay it and gain limited access. The issue is low severity (CVSS 3.1 base score 3.1) and the available signals point to low real-world urgency: EPSS is 0.19%, no CISA KEV status is indicated in the supplied data, and no public exploit was identified at time of analysis. A vendor patch is available in 3.2.0.
Drupal's Webform contributed module can retain sensitive submitted form data in the browser cache, leaving it recoverable by a later user of the same browser profile on versions prior to 6.2.12 and 6.3.1. Exploitation requires a low-privileged authenticated user (PR:L) to reach Webform-rendered content whose access-control and cache-control settings are misconfigured, and the assessed vector rates attack complexity as high with only limited confidentiality impact (CVSS 3.1 base score 3.1). There is no CISA KEV listing and no public exploit identified at time of analysis; EPSS is 0.19% (8th percentile) and the vuln.today priority is monitor, consistent with a low-severity, low-urgency disclosure.
Stored cross-site scripting in the Drupal AI (Artificial Intelligence) contrib module affects versions 0.0.0 through 1.3.13 and 1.4.0 through 1.4.8, allowing an authenticated user with a privileged Drupal account to store input that is later rendered to another user. The vendor CVSS 3.1 vector (AV:N/AC:H/PR:H/UI:R/S:U/C:L/I:L/A:N, base 3.1) requires victim interaction and a high-privilege account, keeping impact to low confidentiality and integrity effects. No public exploit was identified at time of analysis, and a vendor patch is available via Drupal SA-CONTRIB-2026-119.
Authlib's is_secure_transport() TLS guard can be satisfied by a plain-HTTP URL whose userinfo component merely looks like a loopback host, so crafted values such as http://localhost:80@attacker.example/cb are treated as secure even though the parsed host is an external system. The flaw affects authlib versions 0.13 through 1.7.2 and matters where the check is trusted to reject externally supplied URLs - notably OIDC dynamic client registration combined with RP-Initiated Logout (where a public client can register a cleartext non-loopback post_logout_redirect_uris target) and relying parties that call AuthorizationServerMetadata.validate() on an attacker-influenced discovery document. Per the assessed vector (AV:N/AC:H/PR:L/UI:R) the outcome is limited to cleartext redirection and acceptance of cleartext endpoints with low confidentiality impact, requires configured features and end-user logout interaction, and does not bypass OAuth redirect-URI exact matching or forge tokens; a public PoC reference exists in the project's advisory, but it is not verified as working end-to-end exploit code, and there is no CISA KEV entry - the decision label is monitor.
Authorization bypass in SmileCDR HAPI FHIR servers that have REST-hook subscriptions enabled allows an authenticated user holding subscription-creation permissions to query and read FHIR resources they would not otherwise be authorized to access. The exposure is confidentiality-only (CVSS 3.1 base score 3.1, AV:N/AC:H/PR:L/UI:N/S:U/C:L/I:N/A:N): no data modification or availability impact is indicated, and successful exploitation depends on a specific combination of deployment configuration and delegated permission. No public exploit has been identified at time of analysis, and the issue is not in CISA KEV; the priority signal is monitor, consistent with a genuinely low-priority, prerequisite-heavy flaw.
Jitsu Console before 2.14.0-beta.41 allows an authenticated user who knows another workspace's sync ID to read, overwrite, or delete that workspace's stream cursors, causing data re-ingestion or skipped data. The flaw is a missing await on a Prisma findFirst ownership lookup in the sources/state API, so the authorization comparison evaluates a truthy Promise and never rejects the cross-workspace request. It is low severity (CVSS 4.0 score 2.3) and scoped to cursor state, and no public exploit was identified at time of analysis.
Cleartext transmission of credentials and authentication tokens in the HCL AppScan Azure Boards Plugin occurs when the plugin is configured to communicate with the AppScan server over HTTP instead of HTTPS, allowing an attacker who can intercept that traffic to read sensitive data in transit. Deployments configured to use HTTPS are not exposed, and the flaw causes only confidentiality impact with no integrity or availability effect. No public exploit identified at time of analysis; the issue is low severity (CVSS 2.2) and the decision signal is monitor.
massCode 5.10 exposes its Local HTTP Service through a permissive CORS configuration and without API-token authentication, letting a malicious web page loaded in the victim's browser read snippet data from the loopback service. Exploitation requires the local service to be enabled and the victim to visit attacker-controlled content (UI:P), and impact is limited to low-sensitivity information disclosure with no integrity or availability effect (assessed CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:N/A:N). Publicly available exploit code exists, but there is no indication of active exploitation (not CISA KEV), and the vendor addressed it in commit d25d8fa3f5a1e34285c8634f18729650c0c05cda by rebinding to 127.0.0.1, removing CORS, and adding Bearer-token authentication.
OpenBao's Agent and Proxy services before 2.6.3 do not enforce the X-Vault-Header on their quit endpoints, so a request that omits the header can shut down /agent/v1/quit and /proxy/v1/quit even when require_request_header = true is enabled specifically as an SSRF defense. Only deployments that turn that setting on are exposed, and the CVSS vector (PR:H, AT:P) indicates exploitation presupposes a high-privilege position or a server-side request forgery primitive able to reach the listener; impact is availability-only, with no confidentiality or integrity effect and no access to secrets. Vendor-released patches are OpenBao 2.6.3 and 2.7.0, no public exploit code was identified at time of analysis, and the vuln.today priority label is monitor.
KNIME Business Hub before 1.19.0 sets sensitive cookies without protective attributes such as Secure, HttpOnly or SameSite, so session data can be read or intercepted only when a separate cross-site scripting flaw exists in the product or an attacker already holds a man-in-the-middle position on the network path. The issue is rated CVSS 4.0 base 2.1 with Attack Requirements set to Present (AT:P), and it is prioritized as monitor; no public exploit was identified at time of analysis and there is no CISA KEV listing. The vendor states that a workaround is not possible and that upgrading to 1.19.0 is the only complete remediation.
Wasmtime's WASIp1 fd_readdir implementation leaks three bytes of uninitialized host memory per directory entry to guest code that has a preopened directory it can enumerate, disclosed in versions before 36.0.17, 37.0.0 through 48.0.4, and 49.0.0 through 49.0.2. Only guests using the WASIp1 API with filesystem access are affected; WASIp2-or-later guests and guests running without a filesystem are unaffected. The leaked bytes cannot be selected or controlled by the guest, the impact is confidentiality-only, and this is a low-priority issue (vuln.today decision: monitor) with no public exploit identified at time of analysis.
OpenBao 2.5.0 through 2.6.1 will issue a CA-signed certificate bearing an out-of-range IP SAN when a caller submits a CSR to pki/sign/<role> on a role that has allowed_ip_sans_cidr set and use_csr_sans enabled - the latter being the default - so the CIDR restriction that is correctly applied to explicit ip_sans on pki/issue is silently bypassed on the sign path. Only authenticated principals already holding sign permission on such a role can reach the endpoint, and the consequence is limited to integrity (an unauthorized IP embedded in a certificate), which is why the vendor and independent scoring both rate it low severity. The flaw is fixed in OpenBao 2.6.2; no public exploit has been identified at time of analysis and the issue is tracked at monitor priority. Affected deployments are those running PKI roles with CIDR-restricted IP SANs on 2.5.0-2.6.1.
OpenBao's sanitized configuration endpoint (sys/config/state/sanitized) fails to redact the tls_acme_eab_mac_key value, so callers can read the ACME External Account Binding MAC key that operators configured for TLS listener certificate acquisition in versions 2.2.0 through 2.6.3 and 2.7.0. Two preconditions must both hold: the operator must have enabled ACME EAB (a non-default setup, per RFC 8555), and the caller must already hold high privileges (CVSS 3.1 PR:H) sufficient to invoke the administrative endpoint - no user interaction is required. Successful disclosure lets that caller bind their own ACME account to the operator's account at the certificate authority, giving limited confidentiality and integrity impact (C:L/I:L) with no availability effect; no public exploit has been identified at time of analysis and the issue is assessed low priority, consistent with the monitor decision label. Patched releases are OpenBao 2.7.1 and 2.6.4.
OpenBao's AppRole authentication method accepts expired Secret IDs until the once-per-minute tidy sweep removes them, allowing an attacker who already holds a legitimately issued but expired Secret ID to log in beyond its intended TTL. The flaw affects OpenBao releases before 2.6.4 and 2.7.x before 2.7.1; exploitation requires that AppRole is in use, that the attacker possesses such an expired credential (PR:H), and that the narrow window before the next tidy sweep has not yet closed (AC:H), so real-world risk is low. No public exploit was identified at time of analysis, and vendor-released fixes are available in 2.7.1 and 2.6.4.
OpenBao before 2.6.3 silently drops an explicit narrow deny rule when that deny is built on an identity-metadata template containing wildcard or path segments and sits inside a broader explicit wildcard grant, allowing an authenticated principal to reach paths the policy intended to deny. The impact is bounded: access is confined to what the surrounding wildcard already grants, and exploitation needs a specific policy shape plus identity metadata that the attacker can influence - metadata that OpenBao normally receives from a fully trusted IDP. All severity signals converge on low real-world risk (CVSS 4.0 base 2.1; independently assessed 3.1 vector AV:N/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:L), no public exploit has been identified at time of analysis, and vendor patches are available in 2.6.3 and 2.7.0.
OpenBao's unauthenticated root-token generation and rekey endpoints can complete operations without a corresponding response audit record when request audit logging succeeds but response audit logging fails, reducing forensic visibility. OpenBao versions before 2.6.4 and 2.7.x before 2.7.1 are affected, but the impacted endpoints are disabled by default since OpenBao 2.5.3 and must be explicitly enabled for exploitation. No public exploit identified at time of analysis, and the issue is a low-severity logging gap rather than a path to secret disclosure or modification.
OpenBao versions 2.3.0 through 2.6.2 fail to restrict the plugin registration APIs (/v1/sys/plugins/catalog/*) to the root namespace, so a token that already holds write or delete capability on sys/plugins/catalog/* inside a child namespace can mutate, remove, or fabricate plugin registrations through that child namespace's sys/ mount. The result is a denial-of-service condition against plugin catalog integrity rather than code execution: per the vendor advisory, a registration only leads to plugin execution when its command parameter names a real file in a configured plugin_directory and its sha256sum matches that file's digest on disk, which an ordinary tenant cannot satisfy. No public exploit code has been identified at time of analysis and there is no CISA KEV confirmation; the vendor-released fix is OpenBao v2.6.3 and v2.7.0, and this is assessed as a low-priority issue by every available signal.
Sensitive plugin input can be written into OpenBao audit logs when a high-privilege authenticated principal sends malformed TypeKVPair or TypeHeader field values to a plugin, exposing that data to anyone able to read the audit records. Affected releases are OpenBao before 2.6.3 and the 2.7.0 line before the SDK fix in 2.7.1; builtin plugins were fixed in 2.6.3 and 2.7.0. Real-world risk is low because the actor needs high privileges, audit logging must be enabled and its output readable, and no public exploit was identified at time of analysis; the priority label is monitor.
The fzf fuzzy finder (0.36.0 through 0.74.5) builds its masked action string with repeated string concatenation in maskActionContents, giving quadratic runtime in the size of the supplied action text, so a local user can stall the opt-in --listen HTTP server by repeatedly POSTing bodies of roughly 1 MB to the unauthenticated localhost port, blocking the single-threaded server for about 30 seconds per request. Exploitation requires fzf to be running with --listen enabled (a non-default, opt-in feature) and local reachability of the localhost-bound port; the impact is a temporary stall of that server with no confidentiality or integrity effect, and no public exploit was identified at time of analysis. Severity is low (CVSS 4.0 base score 2.0; assessed CVSS 3.1 AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L), and the priority signals converge on routine monitoring rather than urgent action.
Cross-engine type confusion in Wasmtime's Store APIs can produce trampoline signature mismatches, out-of-bounds stack writes, GC heap corruption, or panics when a host embedder passes an object from one Engine to a Store backed by a different Engine. Only embedders that create multiple Engines and directly call affected APIs such as Instance::new, StructRefPre::new, or Tag::new are exposed; guest WebAssembly cannot trigger the flaw and the wasmtime CLI is unaffected. No public exploit has been identified at the time of analysis, and the vendor has released fixes in 24.0.12, 36.0.13, 46.0.2, and 47.0.3.
Denial of service in Wasmtime affects hosts that use the dynamically-typed Val API to lift values from untrusted WebAssembly guests, where crafted guest-returned values can make the host allocate at least about 100x more memory than the hostcall fuel accounting tracks. Versions before 36.0.16, 37.0.0 through 48.0.2, and 49.0.0 are affected, while statically typed bindgen! hosts are not. Vendor fixes are 36.0.16, 48.0.3, and 49.0.1, and the supplied intelligence shows no KEV entry or public exploit identified at time of analysis, with a vuln.today priority of monitor.