Remote command injection in the Shenzhen Aitemi M300 Wi-Fi Repeater (firmware r0-ea7890a) enables unauthenticated network-adjacent attackers to execute arbitrary OS commands by manipulating the enable, name, or mac parameters of the MAC address filter configuration CGI endpoint. The vulnerable sprintf call in the protocol.csp handler passes user-supplied input directly into a shell command string without sanitization. A public proof-of-concept exploit has been published on GitHub, significantly lowering the exploitation barrier, though no CISA KEV listing has been identified at time of analysis.
Stack-based buffer overflow in UTT HiPER 1200GW firmware (versions up to 2.5.3-170306) enables authenticated remote attackers to corrupt stack memory via the EncryptionMode parameter in the PPTP VPN configuration endpoint, potentially achieving arbitrary code execution on the device. The root cause is unsafe use of strcpy() without bounds validation in the /goform/pptpSrvGlobalConfig handler. A public proof-of-concept exploit has been disclosed on GitHub; the vendor did not respond to coordinated disclosure, leaving no official patch available at time of analysis. No public exploit identified at time of analysis as KEV, but POC lowers the practical exploitation barrier significantly.
Full network takeover via authentication bypass in the InfiniteWP Client WordPress plugin before 1.13.6 affects WordPress Multisite installations, where the plugin fails to validate the site-connection state and the authenticity of requests to its remote-management endpoint. Unauthenticated remote attackers can bind their own management key, hijack an administrator session, and gain control of every site in the network, culminating in remote code execution. No public exploit identified at time of analysis; EPSS is low at 0.19% (9th percentile) despite the maximal 9.8 CVSS, and a vendor patch is available.
Missing authorization in code-projects Task Management System 1.0 allows remote unauthenticated attackers to manipulate task records by sending crafted requests to /user/UpdateTaskStatus.php with arbitrary task_id and val parameters. The server performs no permission check before processing the status update, enabling any network-reachable party to alter task data belonging to any user. A public proof-of-concept exploit exists per VulDB and a GitHub issue report, making opportunistic abuse straightforward despite the absence of a CISA KEV listing.
SQL injection in dresende/node-sql-query versions 0.1.25-0.1.28 allows attackers who can influence the joinType option passed to SelectQuery.from() to inject arbitrary SQL into JOIN clauses, because the joinType string was interpolated directly into the query without any allowlist validation. Applications built on this Node.js SQL query-builder library that pass user-controlled input to the joinType parameter are exploitable remotely with no authentication (per the CVSS 4.0 AV:N/PR:N vector). A publicly available exploit exists per the E:P maturity flag and referenced GitHub issue; no confirmed active exploitation (not in CISA KEV). A vendor-released patch is available as v0.1.29.
SQL injection in code-projects Task Management System 1.0 exposes unauthenticated remote attackers a path to manipulate database queries through the task_id parameter in /user/comment_count_user.php. The flaw is a textbook CWE-89 input-neutralization failure in a PHP application with no authentication barrier, and exploit code has been publicly disclosed via GitHub. No CISA KEV listing exists at time of analysis, but the zero-authentication network attack vector combined with available proof-of-concept code significantly lowers the exploitation bar.
Authentication bypass in code-projects Task Management System 1.0 exposes the login endpoint at /index.php to remote unauthenticated attackers who can manipulate the Password parameter to circumvent credential verification entirely. The flaw is classified under CWE-287 (Improper Authentication) and affects all known versions of this PHP application per the wildcard CPE string. A public proof-of-concept exploit is available on GitHub, and no vendor-released patch has been identified at time of analysis.
Unauthenticated remote exploitation of Badaso 3.0.0-alpha's File API allows attackers to bypass permission controls in ApiRequest::class (src/Routes/api.php), resulting in unauthorized read, write, and access operations on file resources. The vulnerability stems from improper permission assignment (CWE-275) in the PHP-based File API route handler. A public exploit has been disclosed via a GitHub issue; the vendor has not responded to the coordinated disclosure.
Command injection in the MSI Radix AXE6600 WiFi 6E gaming router (firmware v781521) lets attackers reach the TelnetSSH function that handles Telnet configuration and inject shell metacharacters, yielding arbitrary command execution as root. Reported to NVD via VulnCheck and tagged as Command Injection, the flaw grants full device takeover, but no public exploit or CISA KEV listing has been identified and no patched firmware version has been published. CVSS 4.0 is scored 9.3 and reflects network reach with total loss of confidentiality, integrity, and availability on the router.
OS command injection in the MSI Radix AXE6600 Wi-Fi 6E gaming router (firmware v781521) lets remote attackers inject shell commands through router configuration functions and gain root on the device. The flaw was reported by VulnCheck, a vendor patch is available via MSI support, and no public exploit has been identified at time of analysis. The vendor CVSS 4.0 score of 9.3 reflects full compromise of confidentiality, integrity, and availability.
SQL injection in code-projects Task Management System 1.0 exposes the pre-authentication admin login endpoint at /admin/AdminLogin.php to unauthenticated remote attackers who can manipulate the email or password parameters to execute arbitrary SQL commands. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) confirms fully unauthenticated, low-complexity network exploitation with no special conditions required. A public proof-of-concept exploit has been published on GitHub, materially lowering the skill threshold for exploitation; however, this CVE does not appear in the CISA KEV catalog at time of analysis.
SQL injection in the WP Directory Kit WordPress plugin before 1.5.5 lets unauthenticated attackers inject arbitrary SQL through an unsanitised, unescaped request parameter, exposing the entire WordPress database. Reported by WPScan and rated CVSS 9.1 (high confidentiality and integrity impact), it affects any site running the vulnerable versions. There is no public exploit identified at time of analysis, and the EPSS score is low (0.18%, 7th percentile), suggesting no observed mass exploitation yet.
SQL injection in the Iptanus File Upload WordPress plugin (all versions before 5.1.8) enables unauthenticated remote attackers to query the underlying WordPress database without any credentials or user interaction. The CVSS scope-change indicator (S:C) reflects that the injection can reach data beyond the plugin's own tables - including WordPress core tables containing admin password hashes, user emails, and site configuration. No public exploit identified at time of analysis, and EPSS at 0.18% (7th percentile) suggests exploitation has not yet been observed at scale, though the unauthenticated attack surface makes it an attractive target as awareness grows.
SQL injection in the CubeWP Framework WordPress plugin (all versions before 1.1.31) allows any authenticated user with Subscriber-level access or above to execute arbitrary SQL queries against the WordPress database via a crafted AJAX action. The vulnerability is compounded by a missing capability check on the affected AJAX endpoint, meaning the lowest-privilege WordPress user role can reach and exploit it without restriction. EPSS is low at 0.16% (5th percentile) and no public exploit or CISA KEV listing has been identified at time of analysis, though the low authentication bar makes mass exploitation feasible on sites with open registration.
Command injection in Tenda CH22 firmware version 1.0.0.1 allows a low-privilege authenticated remote attacker to execute arbitrary OS-level commands by supplying a malicious Name parameter to the formCertListInfo function at the /goform/CertListInfo endpoint. The vulnerability carries a CVSS 4.0 score of 7.4 with high impact across confidentiality, integrity, and availability of the affected device. A public proof-of-concept has been disclosed via a VulDB submission and associated Notion writeup, materially lowering the exploitation bar; no vendor-released patch has been identified at time of analysis.
Path traversal in parisneo/lollms 2.1.0 exposes arbitrary server files to unauthenticated remote attackers via the SPA catch-all route in `backend/routers/ui.py`. URL-encoded dot-dot sequences (`%2e%2e`) bypass Starlette's built-in path normalization and are decoded by Python's `pathlib.resolve()`, allowing directory escape from the static file root with no privileges or user interaction required. No public exploit has been independently confirmed at time of analysis, though the public fix commit at GitHub fully documents the vulnerable code pattern and the bypass technique, substantially lowering the bar for exploit development.
NULL pointer dereference in Crypt::OpenSSL::PKCS12 before version 1.98 for Perl enables process crashes when any caller invokes info_as_hash() on a crafted PKCS#12 file containing a zero-length BMPSTRING bag attribute. The root cause is a zero-size Renew() call that Perl's memory allocator implements as a free() returning NULL, which is subsequently dereferenced inside strlen() within a call to newSVpvn(). With no public exploit identified at time of analysis and an EPSS of 0.19%, active exploitation is currently unlikely, though any service that parses attacker-supplied PKCS#12 data and calls info_as_hash() is directly exposed to denial of service.
Unauthenticated denial-of-service in WP Maps Pro WordPress plugin before 6.1.3 allows any remote attacker without credentials to exhaust server resources by triggering uncontrolled recursion through an exposed AJAX endpoint. The plugin fails to perform a capability check on at least one AJAX action accessible via the wp-ajax-nopriv_ hook and does not restrict the operation that action dispatches. No public exploit code has been identified at time of analysis and EPSS exploitation probability is very low (0.14%, 4th percentile), though successful exploitation can fully degrade availability of the affected WordPress installation.
WPC Order Tip for WooCommerce plugin (all versions before 3.3.1) exposes sensitive customer order data to unauthenticated remote attackers via an unprotected reporting feature that performs no authorization or nonce verification. Any unauthenticated party with network access to the WordPress store can query the endpoint and retrieve billing names, order IDs, statuses, fee amounts, and order dates for any customer. No public exploit or active exploitation (CISA KEV) has been identified at time of analysis; EPSS sits at the 4th percentile, indicating low current exploitation activity despite a high CVSS score.
Unauthenticated column enumeration in the WP Data Access WordPress plugin (all versions before 5.5.79) allows remote attackers to read arbitrary database column values from any table bound to a public front-end form, with the most critical exposure occurring when that table is the WordPress users table - enabling retrieval of all stored password hashes. The CVSS vector (AV:N/AC:L/PR:N/UI:N) confirms zero-barrier network exploitation with no authentication required. No public exploit has been identified at time of analysis, and the EPSS score of 0.14% (4th percentile) indicates negligible observed exploitation activity, though the attack surface is straightforward once a vulnerable form is located.
GeoDirectory WordPress plugin versions before 2.8.169 leak the title and exact geographic coordinates of non-public (pending or draft) listings to unauthenticated HTTP requests due to a missing authorization check on the map marker data endpoint. Any internet-accessible WordPress site running an affected version inadvertently exposes private listing location data to anonymous callers. No public exploit code has been identified at time of analysis, and the vendor has released a fix in version 2.8.169.
Weak password recovery in macrozheng mall (commit 0504e86) exposes the /sso/getAuthCode endpoint of the mall-portal module to remote manipulation, enabling account takeover via a broken authentication flow. The flaw maps to CWE-640 and requires no authentication, though exploitation is rated high complexity. A publicly available exploit has been published; the vendor deleted the corresponding GitHub issue without explanation and has not responded to private disclosure.
Local file inclusion in WP Maps Pro (all versions before 6.1.3) allows unauthenticated remote attackers to include and execute arbitrary existing PHP files on the server by exploiting an AJAX action that lacks both a capability check and path sanitization. Any internet-exposed WordPress installation running the affected plugin is susceptible with no authentication, user interaction, or special configuration required. No public exploit identified at time of analysis and EPSS is low (0.32%), but the zero-barrier entry point and code execution potential elevate the real-world priority well above what the CVSS 6.5 score implies.
Missing authorization and ownership checks in the PiWeb Cancel order / Refund request for WooCommerce WordPress plugin (versions before 1.3.4.34) expose two distinct attack surfaces: unauthenticated order content disclosure and unauthenticated cart manipulation against logged-in users. Any unauthenticated actor can query another customer's order contents by supplying an arbitrary order ID, and can clear and repopulate a logged-in user's cart via a crafted request, degrading shopping session integrity without any privilege. No public exploit has been identified at time of analysis, and the EPSS score of 0.18% (8th percentile) indicates low observed exploitation probability.
Missing authorization on a public REST API route in the Create WordPress plugin before 2.5.4 exposes unpublished content to unauthenticated remote attackers and permanently publishes it as a side effect of the read request. Any WordPress site running an affected version of this plugin may have draft, scheduled, or private posts silently promoted to public visibility without any author or administrator action. No public exploit code has been identified at time of analysis, and EPSS at 0.14% (3rd percentile) signals low current exploitation interest, though the trivially low attack complexity makes opportunistic targeting straightforward once the endpoint is known.
Unauthenticated remote access to unpublished WordPress content is possible through a missing authorization check in the Create plugin (all versions before 2.5.4), with the compounding side effect that accessing the vulnerable REST API route also forces the requested content to be published publicly. Both the confidentiality breach (reading private drafts) and the integrity violation (forced publication) occur in a single unauthenticated request, making this a dual-impact flaw affecting any WordPress site with the plugin installed below version 2.5.4. No public exploit has been identified at time of analysis, and EPSS is low at 0.14% (3rd percentile), though the zero-barrier network vector warrants prompt patching on content-sensitive sites.
Server-side request forgery in dmitriiweb article-scraper-mcp 1.0.0 allows low-privileged remote attackers to manipulate the `url` argument of the `fetch_article` function in `news_scraper_mcp/server.py`, causing the server to issue arbitrary HTTP requests to internal or external resources. Publicly available exploit code exists via GitHub issue #3, and the maintainer has not responded to responsible disclosure, meaning no patch is currently available. Risk is most significant in cloud-hosted deployments where the server can reach instance metadata services or internal network segments.
SQL injection in itsourcecode Hospital Management System 1.0 exposes authenticated attackers to database manipulation via the `delid` parameter in `/viewdoctorconsultancycharge.php`. The vulnerability is PHP-based, requires only low-privilege authentication, and carries a public proof-of-concept exploit documented on GitHub. No active exploitation has been confirmed via CISA KEV, and the CVSS 4.0 score of 2.1 reflects constrained impact, though the sensitivity of healthcare data elevates organizational risk beyond the raw score.
SQL injection in itsourcecode Hospital Management System 1.0 exposes the `/viewdoctor.php` endpoint to database manipulation via the unsanitized `delid` parameter, allowing authenticated remote attackers to read, modify, or potentially delete records including patient and physician data. No active exploitation is confirmed in the CISA KEV catalog, but a public proof-of-concept exploit exists on GitHub, materially lowering the skill threshold for attacks. The CVSS 4.0 score of 5.3 with PR:L reflects a constrained but realistic threat to healthcare deployments with lax access controls.
Stored cross-site scripting in the LWS Optimize WordPress plugin (all versions before 4.1.2) enables unauthenticated remote attackers to inject persistent JavaScript via the plugin's analytics endpoint, which then executes in the browser of any administrator who views the affected dashboard page. The vulnerability requires no authentication to exploit the injection phase, making mass poisoning of sites feasible before any admin interaction occurs. No public exploit is identified at time of analysis; a vendor-released patch is available in version 4.1.2.
Memory exhaustion via decompression bomb in Ash Framework's keyset pagination can terminate Erlang nodes running versions 1.17.0 through 3.31.0. The `decode_values/2` function in `lib/ash/page/keyset.ex` deserializes client-supplied `page[:after]` or `page[:before]` cursors using `:erlang.binary_to_term/2` without rejecting compressed Erlang external term format payloads or bounding the deserialized size, allowing a cursor of a few kilobytes on the wire to inflate to tens of megabytes of heap per request. No active exploitation (CISA KEV) has been identified, but the attack is mechanically trivial for any caller reaching a keyset-paginated endpoint, and the vendor patch test suite includes a working reproduction payload.
Heap out-of-bounds reads and writes in Perl's regular expression engine through version 5.45.1 arise from a signed 32-bit integer overflow when computing the superlinear cache size in S_regmatch. When a subject string of approximately 286,331,153 bytes is matched against a pattern containing 15 or more WHILEM-participating nodes, the bit-count product overflows to 14, producing a two-byte cache allocation that is then indexed far beyond its bounds. A caller passing attacker-controlled input of this size to such a pattern can crash the Perl process or corrupt heap memory; no public exploit has been identified at time of analysis, and the vulnerability is not listed in CISA KEV.
Unsafe deserialization in lmammino/oidc-authorizer (versions up to 0.4.0) exposes confidential data via manipulated JWT claims submitted to the Fixed Message Handler component. The vulnerable code path runs through the `unwrap` function in `src/handler.rs`, which processes the `jwtClaims` argument without adequate deserialization controls - particularly concerning given this component's role as an OIDC authorization gatekeeper. The CVSS 4.0 vector (AV:N/AC:L/PR:N/UI:N) confirms remote, unauthenticated exploitation with low complexity; a public proof-of-concept exists (E:P), and the vendor has not responded to disclosure, meaning no patch has been confirmed.
Denial of service in lmammino oidc-authorizer 0.4.0 can be triggered remotely by submitting a crafted value for the authorization_token argument processed by the parse_token_from_header function in the Authorization Header Parsing component. Any service using this Rust library to validate OIDC tokens via the HTTP Authorization header is exposed to availability disruption without requiring authentication. A public proof-of-concept exploit has been disclosed, though the vulnerability is not confirmed as actively exploited in CISA KEV.
Server-side request forgery in adafap's api-mcp Proxy API Endpoint exposes the server as an open proxy, allowing remote unauthenticated attackers to manipulate the `url` argument passed to the `customAxios` function in `app/api/proxy/route.ts` and force the server to issue arbitrary outbound HTTP requests. All commits up to 92b9a5d04acfec165c7d4ef852496593aa87be06 are affected; the project's rolling release model means no discrete patched version exists. The CVSS 4.0 E:P modifier confirms a proof-of-concept exploit exists, and the vendor has not responded to the coordinated disclosure filed via GitHub issue #4.
Unauthenticated information disclosure in MingSoft MCMS up to version 3.0.6 allows remote attackers to extract sensitive data via the /mdiy/form/get endpoint of the ms-mdiy component. The vulnerability requires no credentials or user interaction and is exploitable over the network, making it trivially accessible to any attacker with HTTP reach to the application. A public exploit has been released (CVSS 4.0 E:P), though the vulnerability is not currently listed in the CISA KEV catalog; the vendor did not respond to pre-disclosure notification, leaving no official patch available at time of analysis.
Information disclosure in MingSoft MCMS through version 3.0.6 exposes sensitive data via the `/mdiy/form/data/list` endpoint in the ms-mdiy component, exploitable by unauthenticated remote attackers with no configuration prerequisites. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) confirms trivial network reachability, and a public exploit has been released (E:P). The vendor did not respond to responsible disclosure, leaving no official patch available at time of analysis.
SQL injection in MingSoft MCMS up to version 3.0.6 exposes the application's database to unauthenticated remote attackers via the formFields argument in the ModelDataImpl.queryDiyFormData function at /mdiy/form/data/list.do. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) confirms no authentication or special conditions are required, making the endpoint directly exploitable over the internet. A public exploit has been disclosed (E:P in CVSS 4.0 supplemental metrics), and the vendor failed to respond to responsible disclosure, leaving users without an official patch.
Missing authorization in Dolibarr ERP's TakePOS module through version 23.0.3 permits authenticated users with only basic cashier-level access to perform sensitive invoice operations - deleting line items, adjusting quantities, modifying prices, and applying discounts - that require the elevated `editlines` permission. The flaw exists because htdocs/takepos/invoice.php processes these actions without verifying the caller holds the `takepos.editlines` right, creating a privilege boundary bypass within the POS subsystem. An upstream patch commit (8992ce8704da947b6abe7b65a6fe59aed736bb81) is available; no CISA KEV listing and no public exploit code identified at time of analysis.
File inclusion in DedeCMS up to 5.7.118 UTF8SP2 allows a remote attacker with low privileges and high-complexity conditions to include arbitrary files via the _4_Setup function within the Installation Wizard (install/index.php). Exploitation results in limited confidentiality, integrity, and availability impact. A proof-of-concept has been publicly disclosed via GitHub, though no active exploitation has been confirmed by CISA KEV. The CVSS 4.0 score of 1.3 reflects the constrained real-world impact and significant attack complexity required.
Server-side request forgery in LosslessCut's experimental built-in HTTP API (versions up to 3.69.0) allows an adjacent-network attacker to abuse the unauthenticated local API by exploiting missing Host and Origin header validation. The missing controls enable DNS rebinding and cross-site request forgery techniques to issue unauthorized API calls through a victim's running LosslessCut instance, yielding limited confidentiality impact. A publicly available proof-of-concept exists per the exploit reference, but no active exploitation has been confirmed - the vulnerability is not listed in CISA KEV, and exploitation is gated behind a non-default experimental CLI flag that must be explicitly enabled by the user.
Improper access control in 3CORESec Trapdoor versions up to and including 1.2.2 allows remote low-privileged attackers to bypass authorization logic via the DefaultFunction handler, yielding limited but real confidentiality, integrity, and availability impacts on the vulnerable system. The 'Authentication Bypass' tag and CWE-284 classification together indicate that the DefaultFunction fails to enforce expected access restrictions, a particularly concerning pattern for a product category whose core purpose is access gating. No public exploit code or active exploitation via CISA KEV has been identified at time of analysis; the CVSS 4.0 score of 5.3 (Medium) reflects low-tier CIA impacts constrained to the vulnerable system itself with no lateral scope change.
SQL injection in lock-upme OPMS allows remote low-privileged attackers to manipulate the `ids` argument within the IN Clause Handler of controllers/messages/message.go, potentially enabling unauthorized read and write access to underlying database contents. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:L/UI:N) confirms network exploitation with low complexity, requiring only a valid low-privileged session, yielding partial confidentiality, integrity, and availability impact. No vendor patch exists as the maintainer did not respond to the coordinated disclosure, and no public exploit or CISA KEV listing has been identified at time of analysis.
Unauthenticated file deletion in WP Photo Album Plus WordPress plugin (all versions before 9.2.07.002) exposes a public REST endpoint that performs no capability or nonce verification, permitting any remote actor to delete generated album export ZIP archives stored by the plugin. The CVSS 3.1 score is 5.3 (Medium) with vector AV:N/AC:L/PR:N/UI:N reflecting zero authentication and trivial access requirements, though impact is scoped only to those regenerable archive files. No public exploit or CISA KEV listing exists; EPSS at 0.14% (4th percentile) confirms negligible current exploitation activity.
Improper IAM permission configuration in NXP GoldVIP's SitewiseCustomFunction Lambda Function Handler exposes AWS resources to remote attackers holding high-privilege credentials in the affected environment. The flaw, tracked as CWE-284, allows authorized high-privileged principals to exceed their intended access boundaries within the AWS deployment - a misconfigured IAM role rather than a missing authentication check. The vendor acknowledged this as a known historical issue resolved beginning with version 1.13.0 and fully addressed in 1.15.0; no public exploit code or active exploitation has been identified at time of analysis.
Improper privilege management in wongcyrus ExcelLexBot through version 0.0.3 allows remote high-privileged attackers to perform unauthorized actions via the ExcelLexBotS3TriggerFunction Lambda handler, resulting in limited confidentiality, integrity, and availability impact. The affected product is end-of-life and will receive no vendor patch, as the maintainer is no longer supporting it. No public exploit code or active exploitation has been identified at time of analysis, though the remote attack vector and absence of a patch increase residual risk for any remaining deployments.
Path traversal in PV-Bhat gemsuite-mcp 1.0.0 exposes arbitrary local files to any authenticated local user who can invoke the tool's Gemini handler functions. The flaw resides in src/handlers/unified-gemini.ts and affects the gemini_search, gemini_reason, gemini_process, and gemini_analyze handlers, all of which accept file_path or file_paths arguments without sufficient sanitization. No patch has been released and the project has not responded to the responsible disclosure; no public exploit or active exploitation has been identified at time of analysis.
Path traversal in Pimzino spec-workflow-mcp versions up to 2.2.6 allows a local low-privileged user to write approval files outside the intended .spec-workflow/approvals directory by supplying crafted values such as '../../../etc/cron.d/evil' as the categoryName argument to ApprovalStorage.createApproval(). The vulnerability is strictly local (AV:L, PR:L) and carries a CVSS 4.0 score of 4.8, indicating limited blast radius. No public exploit code and no CISA KEV listing exist at time of analysis; version 2.2.7, released 2026-05-04, fully remediated the issue.
Solace Extra WordPress plugin before version 1.6.1 exposes an unauthenticated-to-low-privilege write path via an unprotected AJAX action, enabling any subscriber-level user to overwrite post meta on arbitrary posts and deactivate the site's active templates. The dual vulnerability - missing capability check and missing nonce verification - means exploitation is achievable both directly by authenticated subscribers and indirectly via CSRF against any logged-in user, including administrators. No public exploit code or active exploitation has been identified at time of analysis; the vendor has released a corrected version (1.6.1).
Nexter Blocks WordPress plugin before 5.0.2 exposes a REST endpoint for saving global CSS that any Contributor-level user can invoke without proper authorization, causing injected styles to render site-wide for all visitors. The attack surface is limited to sites with untrusted Contributor accounts, and impact is constrained to front-end defacement, content suppression, and UI redressing - not code execution or data exfiltration. No public exploit code or active exploitation has been identified; EPSS of 0.10% at the 1st percentile confirms negligible exploitation probability at this time.