Unauthenticated file disclosure in Open Source Social Network (OSSN) through version 10.1 lets remote attackers download stored album photos and other uploaded files by directly requesting the album/getphoto and album/getcover routes, which perform no authorization check. Because OSSN assigns object and file GUIDs sequentially, an attacker can simply walk the numeric ID space to reach friends-only album photos and files belonging to other users without ever logging in. Publicly available exploit code exists (a HackMD write-up documents the technique), and the issue rates a moderate CVSS of 6.9 (v4.0) with confidentiality-only impact (no modification or denial of service); there is no confirmed active exploitation (CISA KEV). The vendor has published an upstream fix commit, but a released patched version beyond 10.1 has not been independently confirmed.
Malformed keyset acceptance in Tink-Go through 2.8.0 allows unauthenticated remote attackers to bypass structural validation via keyset.NewHandleWithNoSecrets() or ReadWithNoSecrets(), causing malformed public or no-secret keysets (e.g., an enabled key without a valid primary, or a DESTROYED key with no KeyData) to be accepted. This integrity flaw can mislead downstream consumers such as jwt.JWKSetFromPublicKeysetHandle() into producing a JWK set from an invalid handle. Publicly available exploit code exists, but no active exploitation (CISA KEV) has been confirmed; the impact is limited to integrity with no secret disclosure or code execution.
Unauthenticated remote command injection in the end-of-life Edimax BR-6208AC router (firmware 1.32) allows an attacker who can reach the device's HTTP web-management interface to execute arbitrary operating-system commands by manipulating the submit-url argument of the /goform/formAccept endpoint. The assessed CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N, C:H/I:H/A:H) confirms that no credentials and no user interaction are required, and publicly available exploit code exists in the GitHub repository referenced by the disclosure. Edimax was contacted before publication but did not respond and has shipped no patched firmware; the original CVSS 4.0 score of 6.9 understates the confidentiality, integrity and availability impact compared with the independent assessment, and overall exploitability is governed mainly by whether the management interface is reachable from the WAN or only from the local network, making this a moderate real-world priority.
Missing authentication in the remove_announcements.php state-changing endpoint of KusalkaSilva Learning-Management-System lets unauthenticated remote attackers delete announcement records by supplying a chosen ID parameter to a mysql_query call; the CVSS 4.0 base score is 6.9 (CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/E:P), and our independent 3.1 assessment (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:L) confirms the vector is remotely reachable with no privileges and no user interaction, with impact bounded to integrity and availability of announcement data rather than full system compromise. Publicly available exploit code exists (the flaw is documented in the project's own GitHub issue tracker), but there is no indication of confirmed active exploitation (not in CISA KEV). Real-world priority is modest: the target is an obscure, unversioned GitHub learning project with negligible production exposure, and the maintainer has not responded to the report.
An improper authorization flaw in Zoraxy up to and including 3.3.4 lets remote, unauthenticated attackers bypass the Basic Auth, ZorxAuth SSO, and CAPTCHA protections that a reverse-proxy administrator has attached to a protected endpoint. The bypass only works on deployments that have configured a Path Exception rule for one of those auth providers: the exception matcher compares the incoming request path against the configured PathPrefix as a raw string prefix instead of enforcing a path-segment boundary, so any protected URL that textually begins with an exempted prefix is treated as exempt and served without authentication. Publicly available exploit code exists (no CISA KEV listing), the issue is reported via VulDB, and upgrading to 3.3.5-rc1 resolves it.
Payment-amount tampering in macrozheng mall through version 1.0.3 lets a customer-side attacker override the amount submitted to Alipay: the portal's payment endpoint passes the client-supplied total_amount straight into the Alipay bizContent instead of the order's stored payAmount, so an attacker with a valid order can alter what is charged or recorded. Exploitation requires a live Alipay integration and an authenticated customer account holding a legitimate outTradeNo, and impact is confined to order/payment integrity rather than code execution or data exposure (assessed CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N). Public exploit code and the upstream fix commit are both available, but there is no indication the flaw is being actively exploited (not in CISA KEV).
Model-influenced tool calls in the Agno agent framework (agno-agi) through 3.0.6 and 3.1.1 reach Python exec()/runpy()/subprocess without any confirmation gate, so an attacker who can steer the model's input - typically through prompt injection in untrusted data the agent ingests - can cause the agent to run attacker-chosen code on the host or container running the agent. The issue is remotely reachable and unauthenticated (CVSS:4.0 vector AV:N/PR:N/UI:N; our assessed CVSS:3.1 vector is AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H), and publicly available exploit code exists - although no CISA KEV entry indicates confirmed exploitation in the wild at time of analysis. Real-world risk is conditional: only agents built with the PythonTools toolkit that expose the code-executing tools under the vulnerable default (no explicit requires_confirmation_tools) and that accept attacker-influenceable input are exposed.
Missing authentication in Agno (agno-agi) through version 2.7.3 and 3.0.6 causes AgentOS agent endpoints to be reachable by any remote caller when no security_key is configured, because get_authentication_dependency in auth.py applies no authentication check in that default state. Publicly available exploit code exists, and the project was notified through an upstream issue report but has not yet responded, so no vendor-released patch is identified at time of analysis. The flaw is configuration-dependent and is assessed as medium severity (CVSS 4.0 base 5.5, our independent CVSS:3.1 5.3-equivalent vector AV:N/AC:L/PR:N/UI:N with partial C:L/I:L/A:L impact) - deployments that set a security_key are not exposed, so unauthenticated remote access is confined to default-configured, network-reachable instances.
Zen Cart e-commerce stores up to and including version 2.2.2 mishandle attacker-supplied payment data in the PayPal IPN handler, where a crafted txn_id value is processed by zen_db_perform in includes/modules/payment/paypaldp.php and then rendered without proper escaping in the administrator's PayPal payment notification. The vendor's patch replaces the raw output with htmlspecialchars()/ENT_QUOTES escaping across fields in paypal_admin_notification.php, which points to an output-injection (cross-site scripting/stored-injection) defect rather than the SQL-injection class originally reported by VulDB - an unverified framing the independent assessment explicitly flags as contradictory. Publicly available exploit code exists, but exploitation is constrained: the PayPal payment module and its IPN handler must be enabled and reachable so a malicious IPN callback is processed, and the fixed flaw additionally requires an administrator to open the affected notification (user interaction).
SQL injection in Exponent CMS versions up to and including 3.0.2patch2 allows unauthenticated remote attackers to execute arbitrary SQL via the userLogin function in xmlrpc.php when the optional USE_XMLRPC configuration flag is enabled and the endpoint is network-reachable. The flaw stems from unsanitized username input (CWE-89), and publicly available exploit code exists, though exploitation is gated by the non-default XML-RPC setting, limiting real-world exposure. The vendor did not respond to disclosure, so no official fix is available; this is a moderate-risk issue (assessed CVSS 3.1: 6.9) best handled with compensating controls such as disabling USE_XMLRPC or restricting access to xmlrpc.php.
SQL injection in the XCMS password-recovery flow (PublicController::sendFindPwdMail in Admin/Home/Controller/PublicController.class.php) allows a remote unauthenticated attacker to inject arbitrary SQL through the 'email' argument, exposing or altering data in the CMS database. The vulnerable endpoint is the standard find-password mail handler, reachable with no authentication and no user interaction (CVSS vector PR:N/UI:N), which is why the independent assessment rates confidentiality impact High while limiting integrity impact to Low under a CVSS:3.1 base vector of AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:N. Publicly available exploit code exists (a proof-of-concept writeup is linked from the VulDB record), though the flaw is not confirmed actively exploited (CISA KEV) and risk is tempered by the product's very low deployment footprint; no vendor patch has been identified in the supplied data.
Use of default credentials in the PRVTPRO Amnezia Web Panel Backup Download feature (app.py) allows remote attackers to log in with the product's shipped default credential set and download backup data, exposing configuration and any secrets contained in those backups. Affected versions are through 1.6.2 and 1.7.2; publicly available exploit code exists, while there is no confirmed active exploitation (the CVE is not listed in CISA KEV). The CVSS v4.0 score is 6.9 with a vector of AV:N/AC:L/AT:N/PR:N/UI:N and low confidentiality/integrity/availability impact, and the independent assessment agrees the path is remotely reachable with no user interaction and no legitimately provisioned account, limited primarily by whether the default credentials have been changed and whether the panel is reachable from untrusted networks.
Information exposure in code-projects Online Ticket Reservation System 1.0 allows an unauthenticated, remote attacker to fetch the raw database bootstrap file /database/medallion.sql directly over HTTP and read the embedded schema and any seed data it contains. The path is trivial to guess and publicly available exploit code exists (published via a GitHub proof-of-concept referenced from the VulDB entry), though there is no confirmed evidence of active exploitation in the wild and the attack requires the .sql file to sit inside the web-accessible document root with .sql requests not blocked. Impact is limited to confidentiality loss - schema and content disclosure only, with no write access or code execution - so despite a CVSS 4.0 base score of 5.5, the practical business risk is comparatively low.
SHAREit for Android 6.27.38_ww exposes a dangerous JavaScript bridge routine (syncInvoke) that allows attacker-controlled web content rendered inside the app's WebView to invoke privileged native functionality, resulting in limited information disclosure. Exploitation is unauthenticated but requires the victim to open or load a malicious page within the app, as demonstrated by the publicly available '1-click.html' proof-of-concept; no confirmed active exploitation (CISA KEV) is reported. Given the low confidentiality-only impact and the user-interaction prerequisite, this is a moderate, lower-priority issue.
An unauthenticated information disclosure in tsi-coop tsi-dpdp-cms versions 0.5.2 and earlier lets a remote attacker map the internal API parameter schema by sending a valid `_func` value with required fields omitted; the server's JSON schema validator runs before authentication and replies with a 400 error naming every missing field (for example `required property 'fiduciary_id' not found`). The leak is confined to internal parameter/field metadata for the admin, DPO and client APIs (job, API-key, operator and ROPA functions) and does not expose stored data, tokens, or permit writes. Publicly available exploit code exists, and the vendor released a fixed version (v0.5.3) after early coordinated contact; this is a moderate, not critical, priority issue.
Unauthenticated remote attackers can delete arbitrary user accounts in SEMCMS through version 5.0 by hitting the administrative endpoint /admin/SEMCMS_User.php with the parameters Class=Deleted&CF=user, which performs the deletion without enforcing the authorization check that should gate it (CWE-285). Exploitation requires only network reachability to the admin interface - no credentials, no user interaction, and no special configuration - and publicly available exploit code exists (a GitHub proof-of-concept targeting SEMCMS PHP V5.0), though there is no evidence of confirmed active exploitation in the wild. Impact is limited to integrity and availability of user records (account deletion), with no confidentiality, code-execution, or lateral-movement component, and the vendor did not respond to disclosure, so no official fix has been identified.
Unauthenticated SQL injection in the krupalshah EmployeeRecruitmentSystem admin registration handler (source/adminregister.php) allows remote attackers to manipulate backend database queries through crafted input to the registration form. Per the assessed vector (AV:N/AC:L/PR:N/UI:N), no credentials or user interaction are required, so any internet-exposed instance running the project is directly attackable, though impact is bounded to low confidentiality, integrity, and availability loss (CVSS 3.1 base 5.5-style, CVSS 4.0 score 5.5 with E:P). Publicly available exploit code exists (referenced from a GitHub CVE issue and VulDB submission 963381), but this CVE is not confirmed as actively exploited (no CISA KEV listing); EPSS data was not provided. The practical risk is driven less by exploitation difficulty than by exposure: the project is an abandoned, unsupported, rolling-release GitHub repository with no tagged versions.
SQL injection in krupalshah EmployeeRecruitmentSystem (commit 922ff6d and earlier) allows remote unauthenticated attackers to inject arbitrary SQL through the email or password parameter of the login handler at source/loginprocessing.php. The flaw arises from direct use of mysql_query() without parameterization, enabling read or modification of backend database contents. Publicly available exploit code exists, and because the project is an unmaintained rolling release, no vendor patch or supported version is available.
Missing authorization checks in jshERP through 3.6 let any authenticated low-privilege user create tenant-wide master data - settlement accounts, products with initial stock, departments, units, and dictionary entries - because the add and import endpoints authenticate the caller but never verify role or permission. Publicly available exploit code exists in the form of PoC scripts (poc_master_data_write_authorization.py, poc_dictionary_and_config_write.py, poc_global_config_and_enumeration.py, poc_material_dictionary_write.py) that demonstrate the abuse against endpoints such as /account/add, /material/add, and /material/importExcel. The impact is bounded to data integrity within the attacker's own tenant (no confidentiality or availability loss) and requires a valid low-privilege login, making this a moderate-severity broken-access-control issue rather than a critical priority; NVD rates it 5.3 while our independent assessment is consistent with a moderate, integrity-only impact.
Broken access control in TDuck Survey Form through 6.0 allows any authenticated user to inject attacker-chosen fields onto another user's public survey fill page by abusing the batch field-creation endpoint POST /user/form/item/batch/create. Because the server validates ownership only against the first element of the submitted array, an attacker can seed the request with a form they legitimately own and append victim entries keyed by the victim's formKey, which is exposed on the form's public fill page. Publicly available exploit code exists (a Python PoC is published), and exploitation is limited to integrity impact (VI:L) with no confidentiality or availability effect; no confirmed active exploitation (CISA KEV) has been reported.
Authenticated users of jshERP through 3.6 can bypass the depot_flag warehouse restriction to read stock levels, batch numbers, sales figures and priced bill lines from warehouses they are not assigned to, by supplying arbitrary depotId or depotIds parameters to endpoints such as /depotItem/getMaterialDepotStockByParam. Only a low-privileged account inside the target tenant is required (CVSS 3.1 PR:L, read-only C:L impact, no integrity or availability effect), and publicly available exploit code exists, though there is no indication of confirmed active exploitation. Exposure is confined to cross-warehouse data segregation within a single tenant, so single-warehouse or single-operator deployments where no segregation is configured gain no practical advantage from the flaw.
Audit log tampering in jshERP through 3.6 allows any authenticated user, regardless of privilege level, to forge, alter, or erase audit log entries via the /log/add, /log/update, /log/delete, and /log/deleteBatch endpoints. This medium-severity integrity and repudiation issue (CVSS 3.1 base 5.3) lets attackers hide malicious activity or fabricate evidence, and publicly available exploit code exists. Exploitation requires a valid low-privilege session; no admin role or special configuration is needed, but unauthenticated remote exploitation is not possible.
MoguBlog through 6.2 fails to bind network-disk delete operations to the requesting administrator's own account, allowing any authenticated administrator to destroy files and folders belonging to other administrators by submitting their file or folder uid to POST /networkDisk/deleteFile or /networkDisk/batchDeleteFile. The impact is limited to integrity and availability of stored objects (no confidentiality loss) and is constrained by the requirement that the instance has more than one administrator account and that the attacker knows or can enumerate the victim's uid; the exploit is authenticated (PR:L per the assessed vector), requires no user interaction, and works against all three supported storage backends (local disk, Qiniu, MinIO). Publicly available exploit code exists, but there is no indication of confirmed active exploitation, and the severity is moderate rather than critical. A single-administrator deployment has no other admin whose files could be targeted, which materially limits real-world exposure.
jshERP through 3.6 lets any authenticated user whose role is scoped to personal- or department-level data export the complete supplier, customer and member dataset by calling GET /supplier/exportExcel, which omits the role data-scope filter enforced by /supplier/list. The exposed fields include contacts, phone numbers, tax numbers, bank accounts and balances - a broad, tenant-wide confidentiality breach that a low-privileged account can trigger with a single HTTP request; publicly available exploit code exists, and there is no confirmed active exploitation reported. Rated CVSS 4.0 base 5.3 (assessed CVSS 3.1 AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N), this is a medium-severity, read-only information-disclosure flaw: no privilege escalation, code execution, or data modification results, and its practical impact depends on multi-role deployments where data-scope restrictions are actually configured.
An authorization bypass in jshERP through 3.6 lets any authenticated user - even one whose data scope is restricted to personal or department-level records - read other users' purchase, sales, and inventory bills by inserting a purchaseStatus=0 value into the search JSON of a GET /depotHead/list request. The presence of that single parameter causes DepotHeadService.select to omit the creator/owner filter, returning tenant-wide depot head records instead of only the caller's own. Publicly available exploit code exists (a VulnCheck-published PoC script), though there is no KEV-listed active exploitation and the assessed CVSS 3.1 score of 5.3 reflects confidentiality-only impact (C:L, I:N, A:N) requiring a low-privilege authenticated session, making this a moderate-severity information-disclosure issue best addressed by insiders or attackers who already hold credentials.
Authenticated users of jshERP through version 3.6 can create inventory bills that are already in an audited/approved state by submitting status=1 in the body of POST /depotHead/addDepotHeadAndDetail, sidestepping the application's audit permission check and multi-level approval workflow. The flaw is a missing-authorization issue (CWE-862) in which the server trusts a client-supplied status field instead of enforcing approval rights; any low-privilege account able to authenticate and reach the depot-head creation endpoint qualifies, with no victim interaction required. Impact is integrity-only and limited (CVSS 5.3; assessed CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N), affecting stock levels and downstream report queries rather than confidentiality or availability; publicly available exploit code exists (a Python PoC is published on GitHub), but there is no confirmed active exploitation (CISA KEV) at time of analysis, and no vendor patch reference was identified.
Missing authorization in yshop-crm through 2.1.3 lets any authenticated back-office user approve or reject arbitrary invoices through POST /admin-api/crm/invoice/check, bypassing the intended approval workflow by supplying any invoice id. The CrmInvoiceController check handler carries no @PreAuthorize annotation and the approver-identity checks in CrmInvoiceServiceImpl are commented out, so a low-privileged account with a valid session is sufficient; authentication is still required (CVSS v4.0 AV:N/PR:L; internal assessment CVSS v3.1 5.3 AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N), so anonymous remote exploitation is not possible. Publicly available exploit code exists (a Python PoC published on GitHub), but there is no indication of confirmed active exploitation (not listed in CISA KEV). Impact is limited to unauthorized modification of the invoice approval state - no data disclosure, privilege escalation beyond the workflow, or code execution has been demonstrated.
Authenticated low-privilege users of jshERP (through version 3.6) can read any tenant settlement account ledger by supplying an arbitrary accountId to the GET /account/findAccountInOutList endpoint, bypassing personal or department-level data scope restrictions. This missing authorization check (CWE-862) discloses bill numbers, counterparties, amounts, and remarks, but remains confined to the same tenant and does not allow writes or privilege escalation. Publicly available exploit code exists, though active exploitation has not been confirmed and the risk is assessed as moderate severity.
Any authenticated low-privilege user of jshERP through 3.6 can read the tenant-wide operation audit log because the LogController endpoints /log/list and /log/info validate that a session exists but never check the caller's role or ownership of the records (CWE-862 missing authorization). A single valid account, including one with minimal privileges, is therefore enough to harvest other users' login names, client IP addresses, and recorded actions over plain HTTP GET requests, with no user interaction and no non-default configuration required. Publicly available exploit code exists (a PoC script published alongside the request that also references message tampering), but the vulnerability is not confirmed as actively exploited and the assessed impact is limited to information disclosure - no integrity or availability effect - giving a CVSS 4.0 base score of 5.3 and an assessed CVSS 3.1 score of 5.3, a low-to-moderate priority disclosure issue rather than an emergency.
Authenticated administrators can enumerate other administrators' network-disk files on MoguBlog through version 6.2 by sending a POST request to /networkDisk/getFileList, which filters results only by status, file type, or filePath and never by the owning admin_uid. The flaw is an insecure direct object reference (CWE-639) that leaks file names and download URLs, giving a low-privilege admin read-only visibility into peers' stored files, which is most impactful in multi-administrator or shared-tenant deployments. It carries a moderated severity rating (CVSS 5.3, assessed separately as CVSS:3.1 5.3 with only confidentiality impact) and publicly available exploit code exists, though no confirmed active exploitation has been reported.
Broken access control in MoguBlog through 6.2 lets any authenticated administrator move or rename another administrator's network-disk files and folders by sending crafted requests to POST /networkDisk/moveFile, batchMoveFile, or edit, because those handlers update database rows without any admin_uid ownership check. Publicly available exploit code exists (a Python PoC that exercises the missing network-disk ownership check), but exploitation is only meaningful in multi-administrator deployments where the attacker holds a valid admin session and knows or guesses the victim's file paths, making this a low-to-moderate priority issue rather than an emergency. CVSS 5.3 reflects integrity-only impact, with no confidentiality or availability loss and no user interaction required.
Missing authorization checks in yshop-crm through version 2.1.3 allow any authenticated back-office user to approve or reject contract receivables records by calling POST /admin-api/crm/contract-receivables/check, because the controller handler carries no @PreAuthorize annotation and the approver role checks inside CrmContractReceivablesServiceImpl.check are commented out in the shipped code. Attackers can approve receivables by id, inflating the associated contract's return_money value and generating forged approval log entries, giving a low-privilege insider or a compromised back-office account the ability to manipulate financial approval state. Publicly available exploit code exists (a proof-of-concept Python script published on GitHub), although there is no confirmation of active exploitation, and the EPSS/KEV signals in the supplied data do not indicate confirmed in-the-wild use; the practical prerequisite is a valid low-privilege authenticated session (PR:L in the assessed CVSS:3.1 vector).
Any authenticated back-office user of yshop-crm (through 2.1.3) can approve or reject customer contracts without holding approver privileges, because the CrmContractController endpoint POST /admin-api/crm/contract/check ships with no @PreAuthorize annotation and its approver-identity checks commented out. Approving a contract silently mutates downstream deal and purchase data, while rejecting one returns goods to stock - an integrity-only impact with no disclosed confidentiality or availability loss. Publicly available exploit code exists (a PoC script in the cve-request-poc repository), but there is no evidence of confirmed active exploitation; the attack still requires a valid low-privilege back-office session.
Authenticated users of TDuck survey form (through version 6.0) can overwrite other tenants' form themes - colours and image URLs - by submitting a victim's predictable, sequential theme id together with their own formKey to the saveFormTheme handler. The flaw is an insecure direct object reference (CWE-639) with no confidentiality or availability impact; the assessed CVSS 3.1 vector is AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N, so exploitation requires a valid TDuck account and a form that currently has no theme assigned, which is the state in which the handler is reached and the ownership check is skipped. Publicly available exploit code exists (a Python PoC published by VulnCheck's reporter), but there is no confirmed active exploitation (no CISA KEV listing), making this a low-to-moderate, multi-tenant data-tampering risk rather than an account-takeover or data-disclosure issue.
Stored cross-site scripting in Pimcore through 2026.3.1 lets an authenticated backend user who holds data object edit permission plant an attribute-breaking payload in an ExternalImage class-definition field, and that JavaScript then runs in the admin origin when another user opens the affected object's version preview or version diff view. The injected script stays dormant until a second user navigates to the preview/diff screen, so exploitation is not self-triggering and there is no anonymous or default-configuration remote path; publicly available exploit code exists and a technical write-up plus a VulnCheck advisory are referenced. Impact is confined to the admin backend origin and yields limited confidentiality and integrity effects, so this warrants moderate rather than urgent prioritization and should be scheduled with other backend XSS hardening work.
A heap/stack-adjacent buffer overflow in Kamailio's optional sipt module lets remote, unauthenticated attackers crash the SIP server by sending crafted SIP-T/ISUP messages. The flaw lives in isup_get_calling_party() (and the sibling called/redirecting/redirection parses) in src/modules/sipt/ss7_parser.c, where a declared ISUP parameter length is copied into a fixed 26-byte static buffer without being checked against the actual packet length; affected releases are Kamailio up to and including 6.0.7 and 6.1.3. Per independent assessment the realistic impact is availability only (CVSS:3.1 C:N/I:N/A:L), not code execution, and exploitation requires the non-default sipt module to be loaded and its parsing functions invoked from kamailio.cfg on attacker-supplied SIP traffic; vendor-released patches exist in 6.0.8/6.1.4 (commit 3667bebf8d0bea07ed16218fd921d07db87aeb21, PR #4905), and no public exploit was identified at time of analysis.
Stack exhaustion in the ImageMagick MSL decoder affects 6.x before 6.9.13-57 and 7.x before 7.1.2-32, allowing remote, unauthenticated attackers to crash the image-processing process by supplying a set of crafted MSL images whose cross-references drive the decoder's recursion past its stack limit; the result is availability-only denial of service with no confidentiality or integrity impact. Reachability is conditional: MSL is a scripting pseudo-coder rather than a normal raster format, so the target application must actually pass attacker-controlled MSL input to ImageMagick, and MSL is commonly disabled or restricted through the security policy.xml, which both the vendor advisory and independent analysis identify as the primary limiting factor. There is no confirmed active exploitation (CISA KEV) and no public exploit code identified at time of analysis; vendor patches are available as fixed releases 6.9.13-57 and 7.1.2-32.
Repeated processing of attacker-supplied JNG images by ImageMagick before 6.9.13-57, or 7.x before 7.1.2-32, leaves temporary files on disk, letting unauthenticated remote attackers gradually exhaust storage and degrade service availability. The effect is cumulative rather than immediate: a single crafted file leaks only a small amount of disk, so meaningful impact requires many submissions and depends on the host's free space and temporary-directory quotas. Vendor patches exist for both release lines (6.9.13-57 and 7.1.2-32), and no public exploit code or confirmed actively exploited (CISA KEV) status was identified at time of analysis; the CVSS 4.0 score of 6.9 reflects low availability impact only, with no confidentiality or integrity effect.
A memory leak in ImageMagick's SVG encoder (WriteSVGImage) causes allocated memory to be unreleased when crafted images drive the encoder into specific error paths, allowing repeated processing of such images to exhaust available memory and degrade or halt the service. Affected are releases before 6.9.13-57 (6.x) and 7.x before 7.1.2-32, and exploitation requires the target deployment to actually write or convert images to SVG output, not merely read SVG input. The issue is availability-only (no confidentiality or integrity impact), no public exploit code has been identified at time of analysis, and vendor patches are available.
Absolute path traversal in the SiYuan kernel's private plugin HTTP server (SiYuan 3.7.0 through 3.8.5) allows an attacker-supplied kernel plugin to read any file readable by the SiYuan process, including the API token and conf.json, by returning an absolute path in its handler's ResponseFile.Path that the kernel passes unvalidated to gin's Context.File. The plugin's private endpoint requires no authentication or privileges to reach (PR:N), but the CVSS 4.0 vector carries UI:A and our assessment confirms the decisive precondition is user action: a victim must install and load a malicious third-party kernel plugin, so a default installation is not exposed on its own. No public exploit identified at time of analysis and the issue is not in CISA KEV; vendor-released patch is SiYuan 3.8.6, which rejects any path resolving outside the workspace.
Memory exhaustion in SurrealDB 3.1.0 through 3.2.x lets a Surrealism module crash the entire server, denying service to every namespace and database it hosts. The flaw is that host-side buffering of module stdout/stderr output is unbounded and bypasses the configured per-module memory limit, so a module that writes output without newline characters grows the host buffer until memory is depleted. It is confirmed as a privilege-gated issue: an attacker must hold Owner-level rights to install the malicious module (CVSS vector PR:H), and no public exploit code or active exploitation has been identified at time of analysis; instances that do not use Surrealism modules are not exposed.
Malformed XMP metadata profiles in ImageMagick builds from 7.1.1-7 up to (but not including) 7.1.2-32 cause a heap buffer over-read that lets an attacker read past the end of an allocated buffer and typically crash the process. The flaw sits in the XMP profile synchronization code and is reachable by unauthenticated remote attackers (CVSS:4.0 PR:N/UI:N) whenever a service parses an untrusted image - upload handlers, thumbnailers, format converters, PDF/report generators, or any pipeline using an affected library build. Impact is limited to availability (VC:N/VI:N/VA:L, independently assessed as CVSS:3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L), so the realistic outcome is a denial-of-service crash rather than code execution or confirmed data disclosure; no public exploit code or CISA KEV entry has been identified at time of analysis, and a vendor patch is available.
Account enumeration in Cockpit CMS 2.14.1 allows unauthenticated remote attackers to identify valid usernames and email addresses by measuring response-time differences at the /auth/check login endpoint. The flaw is an incomplete timing fix: existing accounts trigger two bcrypt operations while non-existent accounts trigger only one, producing a statistically observable signal. The independent assessment rates this as a genuine-but-moderate information disclosure with no public exploit identified at time of analysis, and an upstream fix commit is available.
Unauthenticated remote attackers can invoke specific Frappe ERPNext endpoints that fail to enforce authorization checks, allowing them to create records and enumerate availability even when an administrator has disabled the corresponding feature. The flaw affects ERPNext before 15.110.0 and 16.x before 16.21.0, is classified as CWE-862 (missing authorization), and carries a moderate risk profile (6.9 CVSS v4.0; independent assessment CVSS 3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:L) with impact bounded to low-integrity and low-availability effects and no data disclosure, code execution, or privilege escalation. No public exploit code or confirmed active exploitation (CISA KEV) was identified at time of analysis; exploitation is nonetheless trivial for any network-reachable instance because no authentication, user interaction, or special conditions are required.
Arbitrary local file read in Craft CMS 4.0.0 through 4.18.6 and 5.0.0 through 5.10.13 lets a user who can set an entry-type title format - directly or by applying project config - abuse the Twig create('DOMDocument').loadHTMLFile() chain to read files from the server filesystem and surface their contents to non-admin users who can view the affected entries. The vulnerability stems from Craft's create() class blocklist omitting the DOMDocument class, and it is limited to confidentiality impact (CVSS 6.9, PR:H, C:H/I:N/A:N) because the attacker must already hold a high-privilege configuration capability rather than anonymous access. No public exploit code or confirmed active exploitation was identified at time of analysis, so this is a moderate-priority information-disclosure issue rather than an emergency, and it matters most in multi-role deployments where the configuring user is distinct from and more privileged than the entry-viewing users who receive the leaked file contents.
ImageMagick releases before 6.9.13-57 and 7.x before 7.1.2-32 expose a single uninitialized byte of process memory through the internal PostScript (PS) decoder, allowing unauthenticated remote attackers who can submit a crafted PS file to read one byte of stale heap contents via the decoded ICC profile. The root cause is an off-by-one in ReadPSInfo in coders/ps.c, where the ICC profile string length is set to the decoded length plus one, so the buffer is one byte longer than the data written into it. Exploitation requires that the application actually routes attacker-supplied PostScript into ImageMagick and that the resulting output (for example the generated image or its ICC profile) is returned to the attacker; no public exploit code or confirmed active exploitation was identified at time of analysis, and the practical impact is capped at a single byte per request with no integrity or availability effect, making this a low-to-moderate priority issue despite the 6.9 CVSS 4.0 base score.
An uninitialized-memory information disclosure in ImageMagick's JPEG decoder affects all 6.x releases before 6.9.13-57 and 7.x releases before 7.1.2-32, allowing a remote, unauthenticated attacker to leak process memory by submitting a crafted JPEG for processing. The flaw is triggered when decoding a JPEG whose component count is fewer than four (for example single-component grayscale or 3-component YCbCr images), causing the decoder to read uninitialized fields from the JPEG component array; the leaked bytes are surfaced through the image's sampling-factor property, so exposure depends on whether the hosting application echoes image metadata back to the attacker. Exploitability is limited and disclosure reachability is conditional; there is no confirmed active exploitation (not in CISA KEV) and no public exploit code identified at time of analysis, though the vendor advisories and fixing commits are public. Vendor-released patches exist in 6.9.13-57 and 7.1.2-32, making this a moderate-priority remediation item rather than an emergency.
ImageMagick's Truevision TGA image decoder can copy uninitialized stack memory into processed output, allowing a crafted TGA file to disclose residual contents of the ImageMagick process's memory to the party that receives or observes the decoded image. Versions before 6.9.13-57 (6.x) and before 7.1.2-32 (7.x) are affected, and exploitation requires a pipeline that automatically passes attacker-supplied TGA input to the decoder (upload, conversion, or thumbnail generation); the CVSS 3.1 vector AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N rates this as a moderate-severity information disclosure with limited confidentiality impact rather than a high-priority compromise. There is no confirmed active exploitation (CISA KEV) and no public exploit identified at time of analysis, and vendor patches that resolve the underlying uninitialized-read defect are available. The practical severity is constrained because the leaked bytes are non-deterministic uninitialized stack data and the attacker must be able to retrieve the decoder's output to recover them.
ImageMagick's CIP (Cisco IP Phone) encoder fails to escape input-derived data, allowing markup injection into the generated XML output when a crafted image is converted to the .cip format. Affected versions are ImageMagick 6.x before 6.9.13-57 and 7.x before 7.1.2-32; exploitation requires that an application actually invoke the CIP coder, which is non-default and must be explicitly requested. Per the assessed vector, the impact is limited to low integrity (markup corruption of the emitted XML) with no confidentiality or availability loss, and no public exploit has been identified at time of analysis (not listed in CISA KEV).
Hard-coded cryptographic keys shipped inside ZKTeco BioTime 9.5.6 Build 20260130.3206 can be recovered by anyone with access to the application binary, installer, or firmware, and then reused against a network-exposed BioTime service with no authentication and no user interaction required (CVSS:3.1 AV:N/AC:L/PR:N/UI:N, Low confidentiality/integrity/availability impact). The vendor-reported CVSS 4.0 base score is 6.9, the disclosure originates from a single source (VulDB), and there is no CISA KEV listing and no public exploit identified at time of analysis; because the report does not identify which feature or data the key protects, realistic impact cannot be confirmed and only version 9.5.6 Build 20260130.3206 is named as affected. Attackers must still be able to obtain the binary/firmware containing the key and reach a BioTime service that actually relies on it before any of the Low-severity impact materializes.