Element X iOS releases before 26.05.0 parse io.element.call:// deep links without validating the embedded target against a URL allowlist, allowing a crafted link to load an attacker-controlled origin into the in-app call screen's WKWebView. Because that web view runs inside the Element X process, it inherits any microphone and camera permissions the user previously granted the app, so the attacker's page can reach device media capture. Exploitation requires the victim to actually open the malicious link and to have already granted Element X iOS media permissions; the vendor fixed the issue in 26.05.0, and no public exploit code or confirmed active exploitation was identified at time of analysis.
Stored cross-site scripting in SourceCodester Inventory System 1.0 allows an authenticated user to inject script through the Hsn_Code parameter of /api/transactions_handler.php?action=save_sale, which executes in another user's browser when the saved sale is viewed. Exploitation requires an authenticated account and victim interaction, limiting real-world urgency; no public exploit identified at time of analysis and no CISA KEV active-exploitation signal was provided. The assessed impact is a limited integrity effect in the victim's browser context, with no confidentiality or availability impact.
Path traversal in GrandNode2 through 2.3.0 lets an authenticated administrator with access to the Excel product-import feature read files from the server by supplying local filesystem paths in the Picture1, Picture2, or Picture3 columns; the vulnerable code passes those values to file-read operations without http(s) or host validation. The issue is medium severity, high-privilege gated, and fixed in 2.4.0; no public exploit identified at time of analysis.
Address-bar spoofing in Sleipnir Mobile for Android before 3.9.0 lets attacker-controlled web pages paint a counterfeit trusted-origin address bar over their own content, because entering fullscreen through the HTML Fullscreen API fails to display the persistent fullscreen notification or trusted-origin indicator users depend on. The forged browser chrome can masquerade as a credential prompt, forced update, or payment page, so the realistic outcome is phishing-driven credential theft, malware delivery, or social-engineering fraud rather than direct code execution or data access. Exploitation is interaction-gated (the victim must open hostile content in this specific browser and complete a single tap), remote attackers need no authentication per the CVSS 4.0 vector, no public exploit code has been identified at time of analysis, and Fenrir has released 3.9.0 to correct the issue.
Stored cross-site scripting in Koalix CRM through 1.14.0 allows authenticated staff users who can create or edit tasks to persist HTML or JavaScript in a task title, which the Task.link_to_task() helper renders unescaped inside the Django admin task changelist and change form. The payload stays inert until an administrator views the affected list or record, so successful abuse requires admin interaction and, at minimum, a valid staff-level account - meaning the realistic threat is a malicious or compromised insider rather than an anonymous remote attacker. Impact is rated medium (CVSS 4.0 base 5.1; assessed CVSS 3.1 AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:N) and no public exploit code or confirmed active exploitation was identified at time of analysis; an attacker who does land the payload could perform admin-context actions such as creating a superuser account.
Frappe Helpdesk releases before 1.26.0 trust a client-supplied raised_by value in the hd_ticket.api.new endpoint, letting an authenticated low-privilege portal user create tickets that appear to originate from another customer's email address. Because that same identity field drives the ticket activity lookups, the attacker can also pull the victim's thread replies through get_ticket_activities, enabling targeted phishing planted directly inside the victim's support portal. Successful abuse requires a valid portal account and passive victim engagement with the fabricated ticket, and no public exploit code or active exploitation has been identified at time of analysis; the vendor-released patch is v1.26.0.
Cleartext storage of remote log server credentials, truststore passwords, and keystore passwords in WSO2 API Manager and related WSO2 products lets high-privileged authenticated Management Console users or anyone with direct read access to product files recover secrets when remote logging is enabled. The flaw is medium severity (CVSS 4.9, CWE-312) and limited to confidentiality; it can enable manipulation of log entries and compromise of the remote log server after credentials are obtained. No public exploit was identified at time of analysis, and deployments without remote logging enabled are not exposed through this path.
Stored cross-site scripting in GrandNode 2 through 2.3.0 lets a back-office operator with product-editing rights (store manager or vendor) plant script inside a product description that Grand.Web later renders unencoded through Html.Raw in Product.Layout.Simple.cshtml, so the payload executes in the browser of any other user who opens that product page. Because the CVSS vector carries PR:H, anonymous or unprivileged attackers cannot inject the payload themselves; the practical exposure is confined to deployments with multiple partially trusted operators, such as multi-vendor marketplaces, and execution still requires a victim to view the compromised page. A vendor fix is available in 2.4.0-beta and 2.4.0 (commit 41a8e349eab8d9641cf6ea800d738f2d8991ca22), no public exploit code has been identified at time of analysis, and this is best treated as a genuine but low-to-moderate priority flaw rather than an urgent threat.
Out-of-bounds read in libaribcaption 1.1.0 through 1.1.2 lets crafted ARIB STD-B24 captions drive DecoderImpl::HandleGLGR() past the 78-entry kLatinSpecialTable, leaking adjacent read-only data or crashing the decoder. The issue is local and requires the victim to process an attacker-supplied caption stream (UI:P, PR:N), with limited confidentiality and availability impact; no public exploit identified at time of analysis. It is fixed upstream in commit 9917cb3.
Apache Camel's camel-http OAuth2 client-credentials token cache is keyed only on the request URI, client ID and client secret, so two endpoints in the same JVM that differ solely in OAuth2 scope, resource indicator or token endpoint collide on one cache entry and one route can be served a token minted for another. Affected are Apache Camel 4.0.0 through 4.18.4 and 4.19.0 through 4.22.0 with token caching enabled (oauth2CacheTokens=true) and at least two such OAuth2 endpoints configured; deployments with a single endpoint, or endpoints that also differ in URI or credentials, are not impacted. Rated CVSS 4.8 (AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N, CWE-524), the practical consequence is authority confusion between the operator's own routes rather than an external attack path, and no public exploit identified at time of analysis; fixed in 4.23.0, 4.18.5 and 4.22.1.
Prefix confusion in Apache Camel's camel-platform-http component silently promotes ordinary HTTP endpoints to forwarding proxies, letting remote unauthenticated callers relay requests to targets they choose via the Host header and reach content the route author never intended to expose. Only Camel deployments running a platform-http route whose configured path begins with the five characters 'proxy' (for example proxyStats, proxy-health or proxying) while not intending the documented platform-http:proxy endpoint are affected - a developer-introduced naming condition that keeps attack complexity high and drives the modest CVSS 3.1 score of 4.8 (AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N). Default deployments without such endpoints are not exploitable, and no public exploit code or CISA KEV entry was identified at time of analysis.
Apache Camel 4.0.0 through 4.18.4 and 4.19.0 through 4.22.0 let a remote sender choose which saga their exchange joins by supplying the unprefixed 'Long-Running-Action' header, because the saga processor consulted that header on every exchange rather than only for saga services that participate in an external LRA protocol. The consequence is limited to deployments running sagas through the camel-lra component with a real LRA coordinator and a consumer that accepts inbound messages from untrusted senders: an attacker can attach a request to a long-running action started by someone else, or force compensation to be associated with an action they did not initiate, yielding low confidentiality and integrity impact (CVSS 4.8, AC:H). Vendor-released patches exist in 4.23.0, 4.18.5 and 4.22.1, and no public exploit has been identified at time of analysis.
Apache Camel's camel-keycloak security processor can accept a token minted by an issuer other than the configured one when token introspection is in use, the opt-in issuer validation is enabled, and the introspection endpoint returns an active response that omits the 'iss' claim - the check was skipped instead of failing. The flaw affects Apache Camel releases from 4.15.0 up to but not including 4.18.5, and from 4.19.0 up to but not including 4.22.1, specifically deployments using the camel-keycloak component with token introspection. Exploitation is unauthenticated per the CVSS vector (PR:N) but demands a non-default chain - introspection enabled, validateIssuer explicitly turned on, and a broker/gateway/minimal introspection endpoint that omits iss and serves multiple issuers - which caps the score at 4.8; no public exploit identified at time of analysis, and vendor-released patches exist in 4.23.0, 4.18.5 and 4.22.1.
Reflected cross-site scripting in MISP workflow module configuration pages affects versions before 2.5.48 and allows an attacker who can influence the parameter names returned by the misp-modules service to inject HTML or JavaScript into form labels viewed by an authenticated MISP user. Exploitation requires influence over the module service plus a victim who actively opens the affected workflow module page, so a crafted link alone is insufficient; the issue is medium severity (CVSS 4.0 base 4.8) and no public exploit was identified at time of analysis. MISP 2.5.48 encodes the generated label with h() and removes the injection path.
A CSS injection flaw in MISP's remote event preview page lets an operator of a linked MISP server write arbitrary CSS declarations into the local UI by embedding a payload in a remote tag's colour value; the injected style is rendered when a local user opens the preview for that remote event, enabling visual spoofing, UI manipulation, and narrow information disclosure but not script execution. Only instances below 2.5.48 are affected, and exploitation requires a pre-existing linked-server trust relationship, tag-creation privileges on the attacker-controlled server (authenticated, low-privileged per the CVSS vector), and deliberate victim interaction with the preview page. The vendor has released a patch (MISP 2.5.48, commit 91323b517) that validates tag colours against a strict hex pattern, and no public exploit code has been identified at time of analysis.
Information disclosure in IBM Concert 1.0 through 3.0.0 can expose sensitive heap-resident data to an authenticated local attacker with low privileges who is able to inspect the application's process memory, with high attack complexity and no public exploit identified at time of analysis. The flaw is CWE-244 (failure to clear heap memory before release), rated CVSS 4.7, and a vendor patch is available. Realistic risk is limited to post-compromise or insider scenarios on hosts where an attacker already has local access.
Missing authorization in openGauss Server 3.0.0 through 6.0.0 lets any authenticated low-privilege database user read raw table-data blocks without holding SELECT on the target table. The user can call pg_relation_filepath() to obtain the relation file path, including compressed _compress paths, then call pg_read_binary_file_blocks() to read blocks because the underlying ReadBinaryFileBlocksFirstCall() and is_row_data_file() checks cover path normalization and data-directory boundaries but not caller privileges. The issue is fixed in 7.0.0-LTS, has a CVSS score of 4.3 (PR:L, C:L), and no public exploit was identified at time of analysis.
Stored cross-site scripting in the WSO2 Carbon management console lets a high-privileged authenticated console user plant malicious JavaScript in event publisher or receiver property values, which executes when another console user opens the affected details page. The flaw requires adjacent-network access, high privileges, and victim interaction and is rated CVSS 4.3, with session hijacking prevented by httpOnly cookies. No public exploit identified at time of analysis and a vendor patch is available; the combined constraints make this a low real-world priority.
Authenticated IBM MQ users holding valid queue-manager credentials can trigger an off-by-one buffer overflow by submitting a specially crafted object name, causing process instability or a crash of MQ components. The impact is availability-only (CVSS 4.3, PR:L, A:L) with no path to data disclosure or code execution, and the CPE data confines affected deployments to containerized environments - IBM MQ Operator and IBM MQ Advanced Container Images - rather than traditional MQ installations. No public exploit code or confirmed active exploitation was identified at time of analysis, and the mandatory authentication step sharply limits exposure to insiders or attackers who have already obtained MQ credentials.
Improper input validation in the Rack Management Controller (RMC) on HPE Compute Scale-up Server 3200/3250 platforms and Superdome Flex servers allows a network-reachable, authenticated low-privilege user to list and modify files on the RMC filesystem (CVSS 3.1: 4.3, AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N). Impact is confined to the management-controller layer - exposure of RMC file contents plus limited integrity damage - rather than code execution or loss of availability on the managed compute partitions. No public exploit code was identified at time of analysis, and practical risk is further reduced because these RMC interfaces are normally deployed on segregated administrative networks.
Malformed compressed message data delivered to containerized IBM MQ deployments (MQ Operator and MQ Advanced Container Images) can drive a minor out-of-bounds read (CWE-125) in the decompression code path, but the broker rejects the malformed request and does not return the over-read bytes, so no message or memory contents are exposed. The attacker must already hold valid MQ connection credentials (CVSS PR:L) and clear high attack complexity (AC:H) by crafting a specific malformed compressed structure, keeping the issue at CVSS 3.1 (Low). No public exploit code or active exploitation has been identified at time of analysis, and IBM has published a vendor fix via its support portal.
Memory and task exhaustion in the russh SSH agent server (versions through 0.64.0) lets an unauthenticated attacker deny SSH agent service by sending a constrained add-key request packed with SSH_AGENT_CONSTRAIN_LIFETIME records. Each 5-byte lifetime record causes the agent to spawn a long-lived tokio task that clones the key blob, producing roughly 141x amplification; the public proof-of-concept in the advisory turned 10.2 MiB of protocol traffic into about 2.1 million live tasks and roughly 1.4 GB of RSS growth. Attackers can reach the agent socket locally or through agent forwarding to an attacker-controlled SSH server, and publicly available exploit code exists, though this is availability-only with no code execution or data disclosure.
Classified Listing, a WordPress classifieds plugin, through version 6.1.5 does not properly validate a specified quantity in input, allowing low-privileged authenticated users to reach functionality that ACLs should restrict and to read limited information (CVSS 4.3, confidentiality-only impact). The flaw is remotely reachable without user interaction but requires a logged-in account, such as one created through open registration commonly enabled on classifieds sites. No public exploit code or CISA KEV entry was identified at time of analysis.
Authorization bypass in the WordPress Event Tickets plugin (all versions through 5.30.1) lets an authenticated user tamper with user-controlled object keys and reach ticketing records that should belong to another attendee, customer, or event. Because exploitation requires a valid logged-in account (CVSS PR:L), the realistic exposure is limited to sites that permit self-registration or that automatically issue accounts to attendees and customers; the impact stays within unauthorized read/modification of ticketing objects (I:L, no confidentiality or availability effect), giving a moderate vendor score of 4.3. No public exploit code has been identified at time of analysis and the issue has not been added to CISA KEV.
Missing authorization in the FluentCommunity WordPress plugin (all versions through 2.11.0) lets an authenticated low-privileged member - for example a self-registered community user on a site with open registration - reach functionality that should be gated behind stricter access controls, resulting in limited information disclosure only (CVSS 4.3, C:L/I:N/A:N). The defect is a CWE-862 missing authorization check that is network-reachable, low-complexity and requires no user interaction, but its moderate rating reflects the prerequisite of an existing account. No public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV.
Apache Camel's Jetty HTTP component (camel-jetty / camel-jetty-common), in versions 4.0.0 through 4.18.4 and 4.19.0 through 4.22.0, installs Jetty's CrossOriginFilter with its stock defaults whenever an operator opts in with enableCORS=true, producing an allow-any-origin policy that also grants credentials instead of the empty policy operators expected to fill in. Unauthenticated remote attackers who lure a victim's browser to an attacker-controlled page (UI:R) can then perform credentialed cross-origin reads against the endpoint and read the responses, provided the victim already holds cookies or session credentials for that endpoint and the endpoint returns sensitive data. Exposure is configuration-dependent because it requires the non-default enableCORS=true opt-in on a Jetty endpoint; no public exploit code or active exploitation has been identified at time of analysis, and fixed versions 4.23.0, 4.18.5 and 4.22.1 have been released by the vendor.
Login CSRF in @fastify/oauth2 7.2.0 through 8.3.1 allows an attacker who has learned a valid OAuth2 state value to complete an authorization flow in a victim's browser and silently sign the victim in to the attacker's account, because state and code-verifier cookies set at login routes with two or more path segments lack a Path attribute and are not cleared by the callback's root-path deletion. The flaw is limited to multi-segment login routes and requires user interaction plus a known state value, is rated CVSS 4.2 with high attack complexity, and is fixed in 8.3.2; no public exploit code or CISA KEV entry was identified at time of analysis.
IBM Concert versions 1.0 through 3.0.0 persist web page content to a local on-disk cache that is not adequately protected, so a second user with an account on the same host can read cached page data belonging to another Concert user. The issue is strictly local (CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N, base score 4.0) and impacts confidentiality only - there is no network attack path, no integrity or availability effect, and single-user or isolated deployments where no other principal can reach the cache are not meaningfully exposed. Exploitation requires an attacker to already hold local/interactive access to the host; no public exploit code has been identified at time of analysis and the issue is not in CISA KEV. A vendor patch is available.
Quadratic processing in markdown-it (prior to 14.3.3 and 15.0.3) allows attacker-controlled Markdown to block the event loop or freeze a browser main thread when HTML parsing is enabled. Only deployments instantiated with the non-default 'html' option or the 'commonmark' preset are exposed; the shipped default disables inline HTML, and HTML blocks are unaffected - the payloads must sit inside a paragraph. Both a public advisory and upstream commits/patched releases exist, but no public exploit code and no confirmed active exploitation were identified at time of analysis, and EPSS data was not supplied; this is a low-priority defect whose severity is capped by the required non-default configuration.
Cross-origin resource sharing (CORS) misconfiguration in IBM Concert 1.0 through 3.0.0 can allow attacker-controlled web content to carry out privileged actions in the context of a victim's authenticated session when the allowed domain is not limited to trusted domains. The vendor rates attack complexity High and impact limited, and the CVSS vector indicates no prior privileges are required (PR:N) even though practical exploitation depends on a victim with an active Concert session loading malicious web content. IBM has released a fix; no public exploit was identified at time of analysis.
Improper certificate validation in IBM Concert 1.0 through 3.0.0 allows an attacker who already holds a man-in-the-middle position to present forged or invalid TLS certificates that the product may accept, yielding low confidentiality and integrity impact with no availability effect. The provided CVSS 3.7 vector requires high attack complexity and user interaction, and the supplied vectors conflict on whether low-level privileges (PR:L) are required, so authentication requirements are not consistently confirmed from the available data; no public exploit code or CISA KEV entry was identified at time of analysis. This is a low-severity issue and the signals are consistent in pointing to limited real-world risk.
IBM Concert 1.0 through 3.0.0 can expose highly sensitive information because it relies on weaker-than-expected cryptographic algorithms, allowing an attacker who obtains the protected data to attempt to recover plaintext. The CVSS vector (AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:N/A:N) describes a network-reachable, unauthenticated and no-interaction path, but exploitation is high-complexity and limited to confidentiality impact. No CISA KEV entry and no public exploit were identified at time of analysis, so this is a low-priority cryptography hardening issue rather than an actively exploited intrusion vector.
Server-side request forgery in alexta69 MeTube (up to 2026.09.28) allows a remote, unauthenticated attacker who can influence a playlist or channel feed to make the instance fetch an arbitrary thumbnail URL - including redirects - from its own network position, writing the internal HTTP response into the download directory that /download serves back. The flaw sits in DownloadQueue.__write_feed_metadata_sync (app/ytdl.py) and is only reachable when metadata/thumbnail writing is enabled (writeinfojson/writethumbnail); exploitation is rated highly complex (AC:H, CVSS 4.0 base 2.9) and impact is limited to internal-response disclosure with no code execution. Publicly disclosed proof-of-concept exploitation details exist (CVSS E:P), and the issue is fixed in release 2026.09.29; it is not listed in CISA KEV, so there is no confirmation of active in-the-wild exploitation.
Path traversal in IBM MQ Web Console allows an authenticated administrator to write files outside the intended directory in deployments using IBM MQ Operator or IBM MQ Advanced Container Images. The issue is low severity (CVSS 2.7) with integrity-only impact and a high privilege prerequisite, and no public exploit identified at time of analysis; a vendor patch is available per IBM advisory.
Cookie host-only protection can be silently stripped from aiohttp's client CookieJar when a client that shares one jar across a parent domain and its subdomains is lured into requesting an attacker-controlled subdomain that returns a same-named cookie carrying an explicit Domain attribute. Only aiohttp releases prior to 3.14.4 are affected, and per the vendor the practical impact is narrow: an attacker gains little unless the application reflects the affected cookies back into a browser context. The issue is low severity (CVSS 2.3, independently assessed as CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:L/I:L/A:N), is not listed in CISA KEV, and no public exploit code was identified at time of analysis; the vendor-released patch is version 3.14.4.
Parse Server fails to enforce class-level create and addField permissions on its built-in _Session class, allowing an authenticated user-or anyone holding that user's valid session token-to create additional persistent sessions and add fields to _Session through POST requests to /sessions or /classes/_Session. Affected releases are Parse Server before 8.6.98 and 9.0.0 before 9.10.3-alpha.1; sessions created this way are not revoked when the original session is logged out. Impact is limited to session integrity with no confidentiality or availability loss, and no public exploit code was identified at time of analysis.
FOSSBilling 0.8.4 through 0.8.5 has an authorization gap where hasManagePermission returns early for inactive non-core modules, letting a low-privileged staff account that holds only module-level extension access read decrypted retained extension secrets via api/admin/extension/config_get and overwrite stored extension configuration via config_save without holding the extension.manage_extensions permission. Exploitation is authenticated (PR:L) and limited to deactivated non-core modules, with low confidentiality and integrity impact and no availability impact, which is why the vendor-assigned and independently assessed severity are both low. There is no public exploit code and no confirmed active exploitation (no public exploit identified at time of analysis); the flaw is resolved in 0.8.6 and is a residual path left over from the earlier CVE-2026-62962 authorization fix.
DOM-based cross-site scripting in the ic-tree-item component of the Intelligence Community UI Kit's @ukic/canary-web-components (pre-release versions 3.0.0-canary.12 through 3.0.0-canary.47) lets unauthenticated remote attackers execute JavaScript in a victim's browser when tree item truncation is enabled and the attacker can influence a tree item's router-item slot label. The flaw is narrow: it only triggers on the non-default truncation code path, requires a victim to render the affected tree, and is limited to low-integrity/low-confidentiality impact within the victim's browser, which is why it carries a low CVSS 4.0 score of 2.3. No public exploit code and no CISA KEV entry were identified at time of analysis; only canary builds are affected, with the fix landing in 3.0.0-canary.48.
Server-side request forgery in smrealms smr's Album Website feature lets a low-privileged authenticated user supply an arbitrary URL in the album 'Website' field, which is passed unvalidated to isUrlReachable() and fetched server-side via curl. The request is headers-only (CURLOPT_NOBODY) with a five-second timeout and the response body is never returned, so this is a blind SSRF that only reveals reachability and HTTP status class, limiting confidentiality impact. Exploitation requires an authenticated account (PR:L per the CVSS vector); publicly available exploit code exists, CVSS 4.0 base score is 5.3 (Medium), and the vendor has released a fix commit but the rolling-release model means no versioned patch is published.
Reflected cross-site scripting in the 'msg' parameter of feedback.php in subhajitkhan's online-clinic-management-system allows remote attackers to inject script into a victim's browser session. The product uses rolling releases, so no specific affected or fixed version is available; exploitation requires a victim to click a crafted link (UI:R) and is limited to integrity impact (no confidentiality or availability impact). Publicly available exploit code exists, but there is no indication of active exploitation (not in CISA KEV), and the issue remains low priority due to required user interaction and niche deployment.
Server-side request forgery in ANX Calibre Manager up to 1.13.12 allows an authenticated low-privileged user to supply a malicious base_url to the LLM Configuration Handler, causing LLMClient.chat to make outbound requests to attacker-chosen destinations. The issue is reachable only when the LLM integration feature is enabled and the attacker can access and submit configuration to that endpoint, with impact limited to low confidentiality disclosure. Publicly available exploit code exists, but there is no CISA KEV confirmation of active exploitation.
Server-side request forgery in TiBillet Lespass v1 API allows an authenticated attacker with high privileges to make the server fetch arbitrary URLs supplied through the img_url or image_url argument in the get_img_from_url function, potentially reaching internal services. The vendor confirms the weakness affects versions up to commit 10e87909c5d526dffa6d424c104fc12860740537 and states that destination IP validation, redirect re-checks, and download timeout/size limits are planned for the v2 API. Exploitation is rated difficult (AC:H) and no public exploit was identified at time of analysis; this is a low-priority, vendor-confirmed issue.
Server-side request forgery filter bypass in Krayin CRM 2.2.5 through 2.2.6 allows authenticated users with webhook permissions to make blind requests to internal services and cloud metadata endpoints via DNS rebinding. The flaw requires winning a time-of-check/time-of-use race: the webhook URL validator resolves a public address, while Guzzle later re-resolves the same hostname to an internal address. No public exploit has been identified at time of analysis, and the assessed severity is low (CVSS 2.1) because exploitation is gated by high privileges and high complexity.
Server-side request forgery in Client St0r through 3.17.586 lets any logged-in user - including read-only accounts - force the server to fetch arbitrary internal URLs through the Property URL Importer endpoint (POST /locations/<id>/import-property-from-url/). Because the pre-patch code only checked for an http:// or https:// prefix and used a plain requests.get() that followed redirects and honoured proxy environment variables, attackers could reach 127.0.0.1, RFC1918 hosts, and the 169.254.169.254 metadata endpoint, with failure exception text echoed back to the requester. Exploit code is publicly available, but no confirmed active exploitation (CISA KEV) is reported; upgrading to 3.17.587 resolves the issue.
Local image file disclosure in NoteGen through 0.38.3 allows authenticated attackers with write access to the victim's configured sync target to plant absolute paths or file:// URLs in synced conversation data, causing the chat image analysis (image_inspect) flow to read local images and send them to the configured AI provider. Exploitation requires the crafted sourceUrl to survive sync merging and the victim to use the image-analysis feature, and impact is limited to confidentiality of local image files. The issue is rated low severity (CVSS 4.0 score 2.1; independently assessed CVSS 3.1 vector with low confidentiality impact), and no public exploit was identified at time of analysis.
SQL injection in iWebShop-5 up to 5.16 via the Util::search function in classes/util.php allows a high-privileged authenticated remote attacker to manipulate database queries, with limited confidentiality, integrity, and availability impact. Publicly available exploit code exists, but active exploitation is not confirmed (no KEV listing provided). The vendor has not responded to the issue report, and the required administrative session significantly limits practical risk.
KMS token exfiltration in Apache Arrow C++ and PyArrow can occur when an application reads an attacker-supplied Parquet file with modular encryption, has a pluggable KmsClient configured, and leaves the KMS URL file-controlled rather than setting it in KmsConnectionConfig. An unauthenticated attacker can exploit this only if the configured KmsClient does not validate the host and actually connects to the file-provided URL, sending a KMS token to an attacker-chosen endpoint; the CVSS 3.1 assessment is AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:N/A:N, with high confidentiality impact and no integrity or availability impact. Affected releases are Apache Arrow C++ 4.0.0 before 26.0.0 and PyArrow 8.0.0 before 26.0.0; no public exploit was identified at time of analysis, and 26.0.0 disables file-controlled KMS URLs by default with opt-in via read_kms_url.
Plack::Middleware::ETag versions from 0.04 before 0.06 for Perl disclose file inode numbers in the default ETag header. When the response body is a filehandle and the file_etag option is not set, the ETag is built from the file's inode number, modification time and size. Clients receiving affected responses may use the disclosed inode numbers to aid backend fingerprinting and correlation of files served through different URLs.
Cpanel::JSON::XS versions from 4.51 before 4.80 for Perl allow quadratic CPU use by revalidating the remaining buffer per value in incr_parse. At the first non-ASCII byte in a string, the decoder validates UTF-8 from there to the end of its input. incr_parse decodes each buffered value with a fresh decoder, so every value revalidates the rest of the buffer, and doubling the input quadruples the work. A peer that sends about a megabyte of ordinary non-ASCII JSON in one read ties up the process for seconds of CPU time.
Cpanel::JSON::XS versions from 4.43 before 4.80 for Perl misorder byte-string keys during canonical encoding in utf16_cmp. canonical sorts keys by UTF-16 code units. When any key in a hash is UTF-8 flagged, every key is decoded as UTF-8 for the comparison, including byte-string keys that the encoder writes out as Latin-1 characters. A byte-string key whose bytes are valid UTF-8 is sorted out of order, and two keys that encode differently can compare equal and are left in hash order. The same data can encode to different canonical documents across processes.