Missing authorization in the Related Posts for WordPress plugin (barrykooij) through version 3.0.0 lets any authenticated Contributor-level user enumerate the titles and publication dates of every private post site-wide by supplying the 'rp4wp_view' parameter to the plugin's admin link screen. The only gate on that screen verifies edit_posts and edit_post against a parent post the attacker already owns - capabilities every Contributor holds - so nothing prevents the private-post query from returning other authors' data. Disclosure is confined to titles and publication dates rather than post bodies or credentials, exploitation requires a valid low-privilege account, and no public exploit code has been identified at time of analysis; the current priority is this_month rather than emergency action.
Missing authorization in the Clean Login plugin for WordPress through version 1.20.1 allows any authenticated user with Contributor-level access or higher to hijack the plugin's front-end page mapping: the save_post-hooked CleanLogin_Shortcode::get_pages_with_shortcodes() routine runs without a current_user_can() check, so a Contributor who embeds the plugin's shortcodes in a post they authored can have the plugin's cl_login_url, cl_register_url, cl_restore_url, cl_edit_url and cl_change_password_url options (plus their matching ID values) overwritten with the ID and permalink of that attacker-owned post. Impact is limited to integrity of those plugin settings - no confidentiality or availability effect and no code execution - and exploitation is impossible without a valid low-privilege account on a site running the plugin. There is no CISA KEV entry and no public exploit identified at time of analysis; the overall priority is medium-to-low, with the practical risk concentrated on sites that hand Contributor accounts to untrusted users.
The Royal MCP WordPress plugin before 1.5.10 exposes non-public content because one of its search tools performs no object-level or post-type authorization checks, so a low-privileged authenticated user such as a subscriber can retrieve the titles and content of private or otherwise unpublished entries; where the companion Royal MCP plugin is installed and stores personal data, that data is disclosable too. The flaw requires a valid WordPress account of any role (PR:L in the assessed vector) and is read-only - no modification, privilege escalation or denial of service - so severity is bounded at 4.3 with EPSS at 0.15% (percentile 4) and no CISA KEV listing; no public exploit identified at time of analysis. It matters most on sites that allow open or self-service registration, where a subscriber account is trivial to obtain, and is scheduled for routine monthly remediation rather than emergency action.
Path traversal in the LangGraph Python SDK (langgraph-sdk 0.4.0 through 0.4.5, and langgraph 1.2.3 through 1.2.13 via RemoteGraph v3 streaming) lets an attacker who controls a streaming thread_id or assistant_id value redirect the SDK's authenticated GET requests to a different readable path on the same configured LangGraph server, with thread_id also able to inject query parameters such as /store/items?namespace=...&key=... . Exploitation requires the application to pass attacker-controlled text directly as those identifiers; requests are GET-only, so the impact is limited to disclosure of another readable path's response (CVSS 4.3, PR:L - authenticated), and the hosted LangGraph Platform is not exposed. The vendor has released fixed versions langgraph-sdk 0.4.6 and langgraph 1.2.14, and no public exploit code was identified at time of analysis.
PostgreSQL Anonymizer before 3.3.0 can leave statistics derived from original, pre-masking data visible in pg_stats after static masking has been applied, until an ANALYZE operation or autovacuum refreshes the table statistics. An authenticated PostgreSQL user with access to pg_stats for the affected relation could read common values, histograms, or range bounds that reveal information about the original data; the exposure is limited to metadata rather than row contents and is time-bounded because routine statistics maintenance clears it. No public exploit identified at time of analysis, and the issue is rated CVSS 4.3 with a fix in 3.3.0.
PostgreSQL Anonymizer before 3.3.0 can disclose entire column values through the anon.partial() and anon.partial_email() masking functions when the underlying data is short enough that partial masking leaves it effectively unredacted. An authenticated database user with query access to an affected masked column can obtain more of the original value than intended, producing a low confidentiality impact (CVSS 4.3, PR:L). The vendor issue states the weakness is fixed in 3.3.0 and later; no active exploitation or public exploit is identified in the supplied intelligence, and the decision label is this_month.
Juniper Networks Session Smart Router can be induced to perform an out-of-bounds read (CWE-125) in its VRRPv3 handling by a crafted VRRP advertisement sent from the same Layer 2 segment, allowing an L2-adjacent attacker to expose stale memory contents or fault the routing process. Affected deployments are those where VRRP is configured and actively running on the targeted interface; the attack requires adjacency on the local broadcast domain but no authentication and no user interaction (CVSS:3.1/AV:A/AC:L/PR:N/UI:N). The confidentiality and availability impact is limited and the trigger specifics are not described in the available sources; no public exploit has been identified at time of analysis, and the vendor advisory JSA112183 is the authoritative reference.
An authenticated user of Xibo CMS can push arbitrary scheduling criteria to any display group through the POST /displaygroup/criteria/{displayGroupId} endpoint, which performs no ownership or permission check on the target group, so the update is relayed over XMR to player devices that should belong to someone else. The result is altered playback - for example weather-dependent content being shown when the condition it depends on is false - but only on installations where the optional Schedule Criteria feature is actually in use, and impact is limited to the integrity of scheduling data (no disclosure, no code execution, no outage). Xibo CMS versions up to and including 4.4.2 are affected and the issue is fixed in 4.4.3; no public exploit code or active exploitation has been reported, and the vendor rates it low severity (CVSS 4.3, PR:L).
SQL injection in Dell PowerProtect Data Domain Management Center lets a low-privileged authenticated user read data they should not be able to access, with no integrity or availability impact. Affected builds are Feature Release 7.7.1.0 through 8.8, LTS2024 7.13.1.0 through 7.13.1.80, LTS2025 8.3.1.0 through 8.3.1.40, and LTS2026 8.6.1.0 through 8.6.1.20. No public exploit code or confirmed active exploitation was identified at time of analysis; this is a moderate-severity, lower-priority issue rather than an urgent emergency, with the main prerequisite being that the attacker already holds valid credentials on the management appliance.
Limited arbitrary media file deletion affects the WordPress plugin AI Engine - MCP Server & AI Chatbot for WordPress, with Claude, ChatGPT & Gemini in all versions up to and including 3.8.4. An authenticated attacker holding only subscriber-level access can abuse insufficient path validation in the plugin's upload_file function to target media files outside the intended upload location, but three prerequisites must align: the administrator must have explicitly enabled the plugin's non-default Public API option, and the plugin's scheduled hourly cleanup_files task must run after the malicious upload before the deletion takes effect. Wordfence rates this Medium (CVSS 4.2, AC:H) with impact limited to integrity and availability of media files and no confidentiality exposure; no public exploit code has been identified at time of analysis, and the issue is not in CISA KEV.
Server-side request forgery in HCL BigFix AEX lets an authenticated low-privilege user induce the application to send requests to external resources, which could expose limited information or enable limited unauthorized actions against reachable services. The flaw is rated CVSS 4.2 with high attack complexity and no availability impact, and the supplied intelligence assigns it a this-month remediation priority. No public exploit code or CISA KEV confirmation is present in the supplied data, and exploitation requires a valid low-privilege session rather than being an unauthenticated default-configuration attack.
Improper control of code generation (CWE-94) in Dell iDRAC Service Module (iSM) versions prior to 6.1.1.0 for Windows, Linux, and ESXi lets a high-privileged local attacker inject code and elevate privileges on the host. Because exploitation requires an attacker to already hold administrator or root-level access on the managed server, the flaw is primarily a post-compromise concern rather than a remotely reachable entry point; CVSS is 4.2 (AV:L/PR:H). No public exploit is identified in the available data and the issue is not in CISA KEV, so remediation should be scheduled through normal patch cycles.
Authenticated SQL injection in the Customer Engagement System WordPress plugin before 2.0.2 lets a WordPress administrator read database contents by supplying unsanitized popup page and category ID values that are concatenated into an INSERT statement. The assessed vector requires a high-privileged administrator (PR:H), which sharply limits the realistic attacker population because administrators already have broad database access, and the issue is rated low priority. EPSS is 0.18% (7th percentile), the vuln.today decision label is this_month, and no public exploit identified at time of analysis.
Authenticated users with only contributor-level access on WordPress sites running the Royal MCP plugin before 1.5.10 can duplicate arbitrary pages by identifier, reading the full content of private and draft pages and republishing them without holding the publish capability. The flaw is a broken object-level authorization issue (CWE-639): the duplication routine relies on a generic capability instead of verifying the caller's rights on the specific page object, so a low-privilege session is enough to cross a trust boundary. Real-world signals are subdued - EPSS is 0.15% (4th percentile), there is no CISA KEV listing, and no public exploit has been identified at time of analysis - and impact is confined to disclosure and unauthorized publication of page content, with no availability or code-execution effect.
Broken access control in the Welcart e-Commerce WordPress plugin before 2.12.6 lets any authenticated user holding the Author role or higher read, modify, create, delete and re-order SKU data - prices, confidential cost prices and stock levels - for products belonging to other users, because the plugin's capability check is not tied to the specific product being edited. Successful abuse requires only a low-privileged network-reachable account and no user interaction, so the confidentiality and integrity impact on store pricing and inventory data is real, but it is bounded: no administrative access is gained and no code execution or outage results. No public exploit has been identified at time of analysis, there is no CISA KEV listing, and EPSS is 0.15% (4th percentile), which aligns with the monitor-level priority assigned to this issue.
Contact-form submissions in Royal Addons for Elementor before 1.7.1070 can be silently redirected to an attacker-controlled email address by any authenticated user holding the WordPress Author role or higher. The plugin stores form email-delivery settings in a site-global namespace keyed only by an author-chosen identifier exposed in public page source, so an Author+ user can read the key from a victim page, overwrite the configuration, and receive that form's submissions. A vendor fix is available in 1.7.1070; EPSS is 0.15% and no public exploit was identified at time of analysis.
Query-parameter injection in i18next-http-backend before 4.0.4 lets attacker-controlled language or namespace values add or override parameters on outgoing translation requests whenever a loadPath or addPath template places {{lng}} or {{ns}} in the query string, such as the multiload shape /locales/resources.json?lng={{lng}}&ns={{ns}}. Server-side deployments using i18next-http-middleware's LanguageDetector are the most exposed case: an unauthenticated visitor supplying ?lng=de%26ns%3Dprivate%26admin%3Dtrue can make the application issue a request to its own translations API as ...?lng=de&ns=private&admin=true&ns=common, leveraging the server's network position and credentials; on the browser side the same crafted link affects the victim's same-origin request. Severity is bounded (CVSS 3.1 3.7, AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:N/A:N) because exploitation requires both a query-shaped template and attacker control of the language or namespace value, and real impact depends on how the translations API branches on the injected parameters; the default /locales/{{lng}}/{{ns}}.json template and supportedLngs-restricted applications are unaffected, the advisory publishes a proof-of-concept script, and no confirmed active exploitation is reported in the available sources (priority label: monitor).
Improper initialization in Juniper Networks Session Smart Router can allow unauthenticated network attackers to cause a limited denial of service, and the description also claims traffic decryption even though the scored confidentiality impact is low. The issue is reached through the router's security policy assignment, and the CVSS AC:H rating indicates conditions beyond the attacker's control; the exact trigger is not described in the available sources. No public exploit was identified at time of analysis, and the vuln.today priority is monitor (CVSS 3.7), making this a low-severity issue rather than an urgent one.
HCL BigFix AEX V13.0 contains a flaw in its bundled aws-sdk package that can cause AWS API requests to be directed to unintended or non-AWS endpoints, with limited confidentiality impact. The issue is rated CVSS 3.7 with high attack complexity and no authentication or user interaction required; the specific exploitation preconditions are not described in the available sources. No public exploit code or active exploitation is identified at time of analysis.
Acronis Cyber Protect 17 on Linux builds before 42587 ship executables that lack complete binary hardening protections (CWE-693), which can make a separate local memory-corruption or exploitation primitive easier to carry out but is not directly exploitable on its own. Abuse requires local access and existing low privileges with high attack complexity, and the resulting impact is limited to low confidentiality and integrity effects with no availability impact. No public exploit identified at time of analysis, and the vendor-released patch is build 42587.
PostgreSQL Anonymizer before 3.3.0 leaves table optimizer statistics fully readable to any role holding SELECT, so a masked role can recover a limited statistical sample of the very values the mask is supposed to hide. Exploitation requires an authenticated principal that has been assigned a masked database role with SELECT on the target table, and yields only histogram bounds and most-common-value entries rather than full rows, which keeps this at low severity (CVSS 3.5, assessed 3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N). No public exploit has been identified at time of analysis and the issue is not in CISA KEV; the fix ships in PostgreSQL Anonymizer 3.3.0.
Masked-role read-only enforcement in PostgreSQL Anonymizer can be bypassed with data-modifying common table expressions (CTEs), allowing an authenticated database user holding a masked role to make unauthorized data changes that the extension is intended to prevent. Affected releases are PostgreSQL Anonymizer versions before 3.3.0; impact is integrity-only, with no confidentiality or availability loss, and deployments that do not use masked roles for read-only enforcement are outside the assessed conditions. The independent assessment rates this as a low-severity, low-priority access-control bypass, no public exploit has been identified at time of analysis, and the complete fix is upgrading to 3.3.0 or later.
An out-of-bounds read in the Bluetooth RFCOMM control-channel handler prh_rfcomm_cc_process_test (inside libanwbtstack.so) on Alpine iLX-F511 in-dash units running firmware 7.0.000 lets a network-adjacent attacker leak limited process memory. The attacker must first pair a malicious Bluetooth device with the head unit - a user-interaction prerequisite (UI:R) - after which a crafted message reaches the vulnerable handler without further authentication (PR:N). Impact of the primitive alone is confined to low-confidentiality disclosure (C:L) for an assessed CVSS of 3.5; the ZDI advisory and CVE description note that chaining with other flaws would be needed to reach arbitrary code execution as root. No public exploit code has been identified at time of analysis, the issue is not in CISA KEV, and the vuln.today decision label is 'monitor'.
Denial of service and limited integrity impact in Dell PowerProtect Data Domain can result from a high-privileged local attacker exploiting a path traversal flaw, with no confidentiality impact (CVSS 3.4). The flaw affects Feature Release 7.7.1.0 through 8.8 and the LTS2024, LTS2025 and LTS2026 ranges listed in the Dell advisory, and requires local access plus high privileges rather than remote or anonymous access. No public exploit was identified at time of analysis, and the issue is low severity with a monitor priority.
Cached OAuth2 introspection responses in WSO2 API Manager, WSO2 Identity Server and related WSO2 identity/gateway products can cause a revoked access token to be returned to the client again until the cache entry is cleared, but only when the allowed_scopes configuration is enabled. An authenticated OAuth2 client able to call both the introspection and revocation endpoints must also hit a narrow ordering window - introspecting a token immediately before revoking it - which is why the issue is rated low (CVSS 3.1, AV:N/AC:H/PR:L, C:N/I:N/A:L). The re-issued token is not accepted for API authorization, so there is no confidentiality or integrity loss; the practical effect is a limited availability/authentication-flow disruption, since affected users may be unable to obtain a new valid token until the cache is flushed. No public exploit code and no active exploitation have been identified at time of analysis, and the priority decision is monitor.
Acronis Cyber Protect 17 builds before 42587 contain an improper authorization flaw (CWE-639) that lets an authenticated low-privilege account read sensitive information outside its permitted scope on Linux and Windows installations. The vendor rates the issue Low severity (CVSS 3.1) with High attack complexity and confined confidentiality impact, and no public exploit was identified at time of analysis; priority is monitor.
The GeoDirectory WordPress plugin before 2.8.195 renders listing data through its field shortcodes without verifying the status or ownership of the referenced listing, so an authenticated user holding the Contributor role or higher can read the content and custom fields of other users' draft, pending, private and trashed listings by supplying the corresponding id attribute. The exposure is read-only - no modification or deletion is possible - and is gated by the requirement for a Contributor-level account on the target site. A vendor patch is available in 2.8.195, EPSS is low at 0.16% (5th percentile) and there is no public exploit or KEV entry identified at time of analysis.
SiteOrigin Page Builder before 2.38.0 exposes unpublished WordPress content to users with contributor-level access or higher through its Post Loop widget. The widget accepts user-supplied query arguments and does not limit results to posts the user is permitted to read, allowing retrieval of other users' draft, pending, scheduled and private posts and pages. This is a low-severity, read-only information disclosure (CVSS 2.7, EPSS 0.16%, priority: monitor); no public exploit was identified at time of analysis and there is no KEV entry.
SQL injection in Dell PowerProtect Data Domain Management Center allows a high-privileged, remotely connected attacker to read information the account would not normally expose; there is no indication of data modification or denial of service. Affected releases span Feature Release 7.7.1.0 through 8.8, LTS2024 7.13.1.0 through 7.13.1.80, LTS2025 8.3.1.0 through 8.3.1.40, and LTS2026 8.6.1.0 through 8.6.1.20. The CVSS 3.1 score is 2.7 (PR:H, C:L), Dell has published a fixed-release advisory, and no public exploit has been identified at time of analysis; the dominant risk factor is the requirement for an existing administrative account.
SQL injection in Dell PowerProtect Data Domain Management Center (DDMC) lets an attacker who already holds high-privilege administrative access read limited data from the management appliance, with no integrity or availability impact. Affected builds span Feature Release 7.7.1.0 through 8.8, LTS2024 7.13.1.0 through 7.13.1.80, LTS2025 8.3.1.0 through 8.3.1.40, and LTS2026 8.6.1.0 through 8.6.1.20; the vector is network-reachable with no user interaction, but the PR:H prerequisite restricts realistic exploitation to insiders or someone who has already compromised an admin credential. At CVSS 2.7 this is a low-severity, low-priority item, no public exploit has been identified at the time of analysis, and the vendor has released fixed builds for each branch.
SQL injection in Dell PowerProtect Data Domain Management Center (DDMC) lets a high-privileged remote administrator disclose limited information from the product's database; the CVSS 2.7 score reflects the requirement for an existing administrative account and only a low confidentiality impact. Affected builds span Feature Release 7.7.1.0 through 8.8 and the listed LTS2024, LTS2025, and LTS2026 ranges; no public exploit or active exploitation is identified at time of analysis, so the issue is rated for monitoring.
SQL injection in Dell PowerProtect Data Domain Management Center (DDMC) allows an attacker who already holds administrative-level remote access to read information they should not be able to reach, with no integrity or availability impact. Affected builds are Feature Release 7.7.1.0 through 8.8, LTS2024 7.13.1.0 through 7.13.1.80, LTS2025 8.3.1.0 through 8.3.1.40, and LTS2026 8.6.1.0 through 8.6.1.20. Scoring is low at CVSS 2.7 (AV:N/AC:L/PR:H/UI:N/S:U/C:L/I:N/A:N) because the PR:H prerequisite sharply limits the attacker pool; there is no CISA KEV listing and no public exploit identified at time of analysis, and the vulnerability is tracked at monitor priority.
The data-shredding workflow in pretix fails to bind generated files to the identity that authenticated the request, because every token-based API user is issued the same session token, so the per-user protection added in 2026.7.1 does not actually take effect. Only authenticated holders of a valid pretix API token can reach the affected code path (CVSS 4.0 PR:L, AC:H, AT:P), and exposure is sharply constrained: shredded files are addressable only through randomly generated UUIDs and are retained for roughly one day before deletion. The vendor rates real-world risk as virtually nil, the issue scores 2.3, no public exploit code has been identified at time of analysis, and the flaw is best characterized as a defense-in-depth regression rather than a directly exploitable data-exposure path. A vendor-released patch (2026.8.1) is available.
Authorization bypass in GROWI v4.5.5 through v8.0.6 lets an authenticated, non-read-only user copy the body of a page revision they cannot read into a page they can edit, via the tags.update API and, under specific editor conflict-check conditions, PUT /_api/v3/page. The attacker must know or guess the target revision's MongoDB ObjectId, which is not secret by design but is not trivially enumerable, making this a low-severity, monitor-level issue. No public exploit was identified at time of analysis, and the vendor fixed the flaw in v8.0.7.
Authenticated read-permission bypass in GROWI versions 3.2.10 through 8.0.6 lets any logged-in user, including read-only accounts without edit permission, read the body or comments of a revision belonging to a page they cannot view by pairing that revision ID with the ID of a page they can view. The flaw is a CWE-639 authorization bypass in the page view, GET /_api/v3/page, and GET /_api/comments.get paths; share-link access is unaffected. It is low severity (CVSS 4.0: 2.3), fixed in v8.0.7, and no public exploit was identified at time of analysis.
MusicFreeDesktop versions 0.0.6 through 0.0.8 expose information through the Request Forwarder component, allowing an attacker on the same local network segment to read limited data without authentication or user interaction. The vendor has not responded to the issue report (GitHub issue #444), and no vendor-released patch or fixed version is identified at time of analysis, while the CVE description states that exploit code has been released publicly. Independent assessment rates real-world risk as low, with confidentiality-only impact and no confirmed active exploitation or KEV listing.
Cross-site scripting in the Vue I18n internationalization plugin (npm vue-i18n, plus @intlify/core and @intlify/core-base) from 9.0.0-beta.5 up to 11.4.13 and in the 12.0.0-alpha line up to alpha.4 allows attacker-supplied markup to reach HTML-rendered translations when escapeParameter (escapeParameterHtml in the Legacy API) is enabled and one translate call supplies both a list and named values. The vendor advisory includes a PoC reference; it is not verified as working exploit code, and no public exploit or active exploitation is recorded at time of analysis (CISA KEV: no; vuln.today decision label: monitor, CVSS 4.0 base 2.1). Impact is limited: three prerequisites must all hold (escapeParameter enabled, translations rendered as HTML such as via v-html, and untrusted named values passed alongside a list), user interaction is required, and the retained sanitizeTranslatedHtml() normally strips event handlers and javascript: URLs, so script execution only follows on builds without the sanitizer or by chaining the iframe srcdoc bypass GHSA-83rm-6c8r-g728; realistic outcomes without that chain are injected phishing links, forms and layout changes.
A remote attacker may be able to terminate a process that passes attacker-controlled input to the P-384 VOPRF API without bounding its length. This vulnerability is addressed in swift-crypto version 5.0.3.
Apache Qpid Proton C through 0.40.0 performs TLS peer hostname verification with its own hand-rolled logic, and a certificate carrying a wildcard entry that does not conform to RFC9525 causes a stack buffer over-read during that check. A malicious TLS peer - or an attacker in a man-in-the-middle position injecting a crafted certificate - can therefore disrupt the connecting client; the assessed impact is availability-only (CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:L), with the user-interaction requirement reflecting that the victim client must actually connect to the attacker-controlled endpoint with verification enabled. Certificates that comply with RFC9525 never trigger the flaw, and the issue is fixed in 0.41.1; no public exploit code was identified at time of analysis and the issue is not in CISA KEV, which is consistent with the low-priority, awaiting-data decision signal.
An authentication bypass in Apache Qpid Proton C through 0.40.0 lets a remote peer step past SASL authentication by sending AMQP performatives that claim to be SASL frames, which the vendor rates as important severity. Impact is conditional rather than automatic: the bypass leaves no user name set, so authentication has not actually completed, and the real outcome depends on whether the hosting application tolerates an identity-less session or falls back to a default identity. No public exploit code or active exploitation was identified at time of analysis, and the vendor has published 0.41.1 as the fixed release tracked under PROTON-2945.
Stack-exhausting unbounded recursion in Apache Qpid Proton C through 0.40.0 lets a deeply nested AMQP message body crash the process when an application decodes it with pn_message_body(). The flaw is availability-only (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H): no code execution or data exposure is described, but any service that passes untrusted AMQP bodies to that decoder can be taken down by a single crafted message. No public exploit code or PoC has been identified at time of analysis, no CISA KEV entry exists, and the vendor has released a fix in 0.41.1.
Apache Qpid Proton C through 0.40.0, when built against OpenSSL, treats the application-supplied trust store as an addition to the host's default system trust anchors rather than as a replacement. Applications that call pn_ssl_domain_set_trusted_ca_db() or pn_tls_config_set_trusted_certs() to narrow trust to a specific CA therefore silently continue to trust every CA in the operating system store, so an attacker positioned to intercept the TLS connection could present a certificate chaining to any of those residual roots and be accepted for the targeted peer identity. The Win32 SChannel backend is unaffected, the scope of abuse is a trust-scope bypass rather than code execution, and an attacker also needs a man-in-the-middle position. No CVSS vector, EPSS score, KEV entry, or public exploit code was provided at time of analysis; the upstream JIRA issue is marked Resolved/Fixed with fix version proton-c-0.41.1.
Hop-by-hop header leakage in Caddy's reverse proxy during WebSocket upgrades allows an upstream that emits headers such as Set-Cookie, Alt-Svc, Keep-Alive or Proxy-Authenticate to have them relayed unmodified to clients instead of being stripped as RFC 7230 §6.1 requires. Caddy v2 releases up to and including 2.11.4 are affected when configured with reverse_proxy and handling HTTP 101 Switching Protocols; the assessed vector (AV:N/AC:H/PR:H/UI:R/S:C/C:L/I:H/A:N) reflects that the attacker must control or compromise the upstream response and that a victim must actually open the WebSocket, with high integrity impact and limited confidentiality impact. A public proof-of-concept reproduction script is included in the GHSA advisory, no KEV entry or EPSS score was provided, and the vuln.today decision label is 'monitor'.
MIME::Field::ParamVal versions before 5.521 for Perl do not flag a duplicate boundary parameter whose name contains a nested comment. A parameter written boundary((c))="A" or boundary(\))="A" is stored under that literal name rather than as boundary. A second boundary parameter is then not reported as a duplicate, and MIME::Parser splits the message on it. Readers that skip comments as RFC 5322 allows take the first boundary instead. A remote sender can hide an attachment from a filter that relies on ambiguous_content() while the recipient's mail client shows it.
HTTP::Body versions before 1.24_01 for Perl allow memory exhaustion via unbounded buffering of chunked request bodies. In chunked mode, used when no content length is given, add() holds each chunk in memory until all of it has arrived. A line that is not a chunk-size line, a size line without CRLF, or a size too large to fit an integer leaves the parser waiting forever, and every later add() grows the buffer. A remote attacker can exhaust memory in applications that pass chunked request bodies to HTTP::Body.
HTTP::Body versions before 1.24_01 for Perl allow CPU exhaustion via regex backtracking on the filename of a multipart upload. For every file part, the multipart parser finds the suffix of the uploaded filename by matching the default $basename_regexp, qr/[^.]+(\.[^\\\/]+)$/, against the whole basename. A basename made of many dots and ending in a backslash makes the regex backtrack in time quadratic in its length. A remote attacker can exhaust CPU with one request, as a basename of about 32 KB takes several seconds to match and every file part is matched separately.
Memory exhaustion in the Apache Kafka broker lets an authenticated principal holding Write permission on a topic submit a roughly 90-byte, CRC-valid compressed produce request that declares a decompressed record body of up to about 2 GiB, forcing the broker to attempt that allocation and raise OutOfMemoryError. The flaw is reachable on the produce-validation path when a batch's codec differs from the topic's configured compression.type and records are re-encoded, and a crafted record accepted at ingest can also kill the single default log-cleaner thread when it is later read on a compacted topic, stopping compaction broker-wide (including __consumer_offsets and __transaction_state) across restarts until the record is removed; because OutOfMemoryError is an Error rather than an Exception, the cleaner thread is not restarted. Impact is availability only - no data disclosure or modification - and no public exploit code has been identified at time of analysis. Critically, patching alone does not close exposure: the new max.decompressed.message.bytes setting defaults to an effectively unlimited (~2 GiB) value for backward compatibility, so operators must both upgrade and configure a bounded limit.
Apache Kafka brokers that have an authorizer enabled will read a file or environment variable from the broker host and return its contents to any authenticated principal, including one holding no ACL grants at all - not even ALTER_CONFIGS. The flaw is a missing-authorization ordering problem (CWE-862): an IncrementalAlterConfigs or AlterConfigs request aimed at a BROKER resource reaches configuration validation, which resolves ${file:/path:key} or ${env:VAR} variables through a ConfigProvider instantiated from the request, before any authorization check is applied, and a type-validation error reflects the resolved secret back to the caller. This defeats the org.apache.kafka.automatic.config.providers allowlist introduced for CVE-2024-31141, but the read is bounded by the broker process's OS file permissions and impacts only confidentiality, not cluster availability or data integrity; there is no public exploit identified at time of analysis and the vendor-supplied fix set is Kafka 4.2.2, 4.3.2, and 4.4.0.
Kafka Connect workers leak plaintext secrets through the REST API's /connector-plugins/{connector-type}/config/validate endpoint, which resolves Configuration Provider variables and returns the resolved values in its response, so anyone who can reach the endpoint can read credentials that were meant to stay hidden until runtime. Deployments are affected from Kafka 2.4.0 up to but not including 4.2.2, and from 4.3.0 up to but not including 4.3.2; the vendor fixed the behaviour in 4.2.2, 4.3.2 and 4.4.0. The request path requires no credentials (the assessed CVSS vector is AV:N/AC:L/PR:N/UI:N, and the Connect REST API has no authentication enabled by default), though disclosure is bounded to secrets that the Configuration Providers actually configured on that worker can resolve; no public exploit code, CISA KEV listing, or EPSS score was identified at time of analysis.
Authentication bypass in Apache Kafka Connect's isolated admin REST API allows unauthenticated network attackers to read and write logger settings, including cluster-wide changes via scope=cluster, when admin.listeners is explicitly configured as a separate listener and REST security extensions such as BasicAuthSecurityRestExtension are used. The issue affects Apache Kafka 2.4.0 through versions before 4.2.2 and 4.3.0 through versions before 4.3.2, with fixes released in 4.2.2, 4.3.2, and 4.4.0; deployments that leave admin.listeners unset are not affected. No public exploit was identified at time of analysis, and the assessed severity is moderate with limited confidentiality, integrity, and availability impact confined to the /admin/loggers endpoint.