Server-side request forgery in MLflow before 3.15.0 lets remote unauthenticated attackers coerce the server into making outbound requests to internal or cloud metadata endpoints via the POST /api/2.0/mlflow/webhooks/{id}/test endpoint. Because URL validation and the actual delivery re-resolve the hostname independently and follow redirects, an attacker can bypass the public-IP check and receive the target's response_status and response_body, enabling exfiltration of cloud credentials (e.g. from 169.254.169.254). No public exploit identified at time of analysis; the CVSS 9.3 rating and scope-change flag reflect the credential-theft potential of reaching metadata services.
Sandbox escape to remote code execution in the vm2 Node.js sandbox library (versions <= 3.11.5) arises because handleException() in lib/setup-sandbox.js sanitizes SuppressedError and AggregateError sub-errors but never inspects the ES2022 Error.cause property. Untrusted code running inside the sandbox that catches a host-thrown error carrying a .cause (or other own property) pointing at a host object such as process can dereference it to reach child_process and run arbitrary host commands. A working proof-of-concept demonstrating full RCE is published in the GitHub advisory, and the flaw is fixed in vm2 3.11.6; there is no CISA KEV listing indicating active exploitation.
Sandbox escape in the vm2 Node.js sandboxing library (npm, versions <= 3.11.5) lets code running inside the sandbox reach host prototype mutators and execute arbitrary OS commands on the host. It bypasses the earlier GHSA-v6mx-mf47-r5wg fix by stacking Function.prototype.call indirection so the dangerous __proto__ setter is no longer in the inspected position, then leaks a live host reference through a rejected WebAssembly.compileStreaming promise to pivot to child_process. A working proof-of-concept is published in the advisory, so publicly available exploit code exists; the flaw is not listed in CISA KEV.
SQL injection in AzuraCast's backup restore flow (versions up to and including 0.23.7) lets an attacker achieve privilege escalation because the `azuracast:restore` command runs the extracted `db.sql` file verbatim, with no content validation. The flaw carries a critical 9.8 CVSS from the reporter and has publicly available exploit code, though its EPSS probability is low (0.23%, 14th percentile) and it is not on CISA KEV. The real-world attack surface hinges on who can supply and trigger restoration of a crafted backup archive.
Remote code execution in Wavlink WN531P3 and WN535M1 (firmware V250922) consumer routers arises from an unsafe strcpy in the export_pingortrace.cgi binary, where the attacker-controlled HTTP_COOKIE header is copied into a fixed stack buffer without bounds checking. Unauthenticated remote attackers can send a crafted Cookie header to the device web interface to overflow the stack and hijack execution. Publicly available exploit code exists (disclosed via VulDB with a PoC report), but there is no public exploit identified as actively exploited in the wild, and no CISA KEV listing.
Authentication bypass in EFM ipTIME A3004T (firmware 14.19.0) lets remote attackers defeat session validation through the httpcon_check_session_url function, granting unauthorized access to the router's web management interface. Publicly available exploit code exists, though the flaw is not yet listed in CISA KEV. The vendor was notified but did not respond, so no fix is currently available.
Unauthenticated data tampering in GitLab CE/EE (18.2-18.11.10, 19.0-19.0.7, 19.1-19.1.5, 19.2-19.2.3) lets a remote attacker modify or delete public projects and user data by abusing a GraphQL directive, per CVE-2026-19478 (CVSS 9.4). The flaw was self-reported by GitLab and publicly available exploit code exists via a HackerOne disclosure, though it is not listed in CISA KEV. Impact is weighted toward integrity and availability rather than data theft (CVSS C:L/I:H/A:H).
Command injection in Glances prior to 4.5.6 allows a local unprivileged user to execute arbitrary shell commands when administrator-configured action templates use adjacent unescaped Mustache variables. The per-field sanitizer in _sanitize_mustache_dict() strips multi-character shell operators from individual template field values but cannot prevent their reconstruction across variable boundaries after Chevron rendering, enabling '&&' to be assembled from two separate '&' characters that each pass the filter individually and are then executed by secure_popen(). No public exploit has been identified at time of analysis, though GHSA-qcpp-8x79-hhp3 provides sufficient technical detail to reproduce the attack technique.
Denial-of-service in vm2 (Node.js sandbox library) versions 3.11.0-3.11.5 allows sandbox code to bypass the `bufferAllocLimit` DoS mitigation via two unwrapped allocation paths - `Buffer.concat(list, totalLength)` and `Buffer.from({length: N})` - committing unbounded host external memory in a single synchronous C++ call that V8's timeout mechanism cannot interrupt. A publicly available PoC confirms that a 200-byte sandbox payload inflates host RSS by hundreds of megabytes per invocation, making a single crafted request sufficient to OOM-kill a host process running inside a Docker container, Kubernetes pod, or AWS Lambda function. No public exploit identified in CISA KEV at time of analysis, though functional exploit code is embedded in the GHSA advisory itself.
Server-Side Request Forgery (SSRF) in 9Router 0.5.4 and earlier allows unauthenticated remote attackers to probe internal network services by supplying an arbitrary issuerUrl to the /api/auth/oidc/test endpoint, which the application fetches without validating against private or loopback address ranges. The reflected OIDC discovery fields - including token_endpoint and jwks_uri - expose internal service metadata, enabling internal network enumeration. No public exploit code or CISA KEV listing has been identified at time of analysis, though the unauthenticated, low-complexity attack surface makes opportunistic scanning plausible.
Stack exhaustion denial-of-service in deepmerge-ts crashes Node.js processes when attacker-controlled recursive object graphs are passed to `deepmerge()`, `deepmergeInto()`, `deepmergeCustom()()`, or `deepmergeIntoCustom()()`. All versions prior to 8.0.0 lack cycle detection in the recursive record-merge routine, allowing two self-referencing objects sharing a key path to drive the process into infinite recursion until the Node.js runtime throws `RangeError: Maximum call stack size exceeded`. A public proof-of-concept is included in the GitHub security advisory; no CISA KEV listing exists at time of analysis.
Memory exhaustion in Netty's SctpMessageCompletionHandler (netty-transport-sctp) allows unauthenticated remote peers to drive a JVM into OutOfMemoryError by sending oversized, incomplete SCTP message fragments. The handler enforces limits on concurrent incomplete messages and per-message fragment counts - introduced as a fix for the prior CVE-2026-46340 - but omits a corresponding cap on total buffered bytes. With default settings of 128 messages and 128 fragments, an attacker can buffer approximately 1GB per connection; a small number of concurrent connections suffices to exhaust server heap. Vendor-released patches exist in versions 4.1.137.Final and 4.2.17.Final; no active exploitation or public POC has been identified at time of analysis.
GraphQL multiplex query handling in GitLab CE/EE improperly validates request methods, permitting unauthenticated attackers to trigger state-changing mutations via HTTP GET requests, effectively bypassing CSRF protections. All GitLab CE/EE deployments running versions 18.2 through the affected ranges across the 18.x, 19.0, 19.1, and 19.2 release lines are exposed. A publicly available exploit exists via HackerOne disclosure, and successful exploitation allows an attacker to perform arbitrary data-modifying operations on behalf of any authenticated GitLab user who can be lured to a malicious page.
MLflow's `CreateModelVersion` endpoint prior to version 3.15.0 permits authenticated users to anchor a model version to another user's run or logged-model artifact directory, then retrieve those artifacts through `GET /model-versions/get-artifact` without holding the required READ permission on the source resource. This confused-deputy pattern arises because `_validate_source_run()` and `_validate_source_model()` in `mlflow/server/handlers.py` enforce only path containment - not cross-user authorization - allowing the registered model's artifact-read gate to be used as a proxy to exfiltrate a victim's ML artifacts. No active exploitation has been confirmed (not in CISA KEV), and the vendor-released fix is available in MLflow 3.15.0 alongside a public security advisory and PR diff that fully document the exploit technique.
Netty's CorsHandler silently overwrites any application-set Vary headers with only 'Vary: Origin', enabling caching proxies and CDNs to conflate and serve authenticated, user-specific HTTP responses to unauthorized requesters. Affected are all Netty deployments running netty-codec-http prior to 4.1.137.Final or 4.2.17.Final where applications set per-user Vary headers (e.g., Vary: Authorization or Vary: Cookie) and are deployed behind a caching layer. No public exploit identified at time of analysis; vendor-released patches are confirmed available in 4.1.137.Final and 4.2.17.Final.
Missing authorization enforcement on the log-inputs API endpoint in MLflow 3.13.0 through 3.14.x allows any authenticated user to inject fabricated DatasetInput records into another user's run lineage metadata without holding UPDATE permission on that run. The root cause is that the LogInputs (and LogOutputs) handler was simply never registered in the BEFORE_REQUEST_HANDLERS auth middleware map, leaving the POST /api/2.0/mlflow/runs/log-inputs endpoint unguarded. No active exploitation has been confirmed (not listed in CISA KEV), and a vendor-released patch is available in version 3.15.0.
Integrity bypass in ApostropheCMS page management allows authenticated low-privileged editors to move pages into restricted subtrees they have no create rights over, and indirectly re-rank admin-only sibling pages via an unchecked database updateMany call. All deployments running ApostropheCMS prior to 4.32.0 are affected; the flaw exists in the REST page update endpoint and requires no special configuration to reach. No public exploit code exists and the vulnerability is not listed in CISA KEV, but the attack path through documented REST parameters is straightforward for any CMS insider.
Stored cross-site scripting in ApostropheCMS prior to 2.17.6 enables execution of arbitrary JavaScript in victim browsers through a parser-differential bypass in the bundled sanitize-html library. When `textarea` or `xmp` is included in a site's `allowedTags` configuration, a payload using a trailing solidus on the closing tag - such as `</textarea/>` - is treated as text by htmlparser2 10.x but as a valid close by spec-compliant browsers, causing injected markup like `<img onerror=...>` to execute as live HTML on page load. No public exploit repository has been identified and the vulnerability is not listed in the CISA KEV catalog, but a functional proof-of-concept payload is fully described in the upstream security advisory GHSA-jxwj-j7wr-gfrw; the fix ships in version 2.17.6.
Remote code execution in Regular Labs Sourcerer for Joomla (versions before 14.0.0) allows unauthenticated attackers to inject and execute arbitrary PHP via {source} blocks. The extension parses {source} tags found anywhere in Joomla's final rendered HTML without reliably verifying where that content originated, so attacker-controlled reflected input reaching the page body gets evaluated as trusted code. Rated CVSS 4.0 10.0 (critical); no public exploit identified at time of analysis, and it is not listed in CISA KEV.
Server-side template injection in ERPNext (Frappe framework) lets low-privileged authenticated users escape the Frappe safe-execution sandbox and achieve remote code execution on the server. The root cause is that frappe.render_template was exposed inside the safe_exec globals without forcing restrict_globals, so a user who can supply template content crosses a permission boundary into unrestricted Jinja evaluation. Rated CVSS 9.9 with a scope change; no public exploit code or CISA KEV listing is identified at time of analysis, but the fix commit is public, narrowing the reversing effort required.
Privilege escalation in Red Hat Advanced Cluster Management's multicloud-operators-subscription controller lets a low-privileged user on a managed cluster create a Subscription with crafted annotations that the controller processes with its own elevated Service Account, deploying attacker-chosen resources into arbitrary namespaces. The flaw (CWE-863, CVSS 9.9, scope-changed) affects Red Hat ACM 2.x and Multicluster Global Hub as reported by Red Hat. There is no public exploit identified at time of analysis and it is not listed in CISA KEV.
Privilege escalation in Red Hat Multicluster Engine for Kubernetes (MCE) lets a privileged service account on a managed spoke cluster submit a forged Certificate Signing Request that the managedcluster-import-controller auto-approves, yielding administrative credentials on the central hub cluster. The controller's auto-approval path neither inspects the CSR signer name nor decodes the PEM-encoded x509 request, so an attacker who already controls a spoke can pivot to full control of the hub and, by extension, every cluster it manages. No public exploit identified at time of analysis; EPSS is low (0.22%, 12th percentile) and CISA SSVC records no observed exploitation.
Server-Side Request Forgery in the WooMS WordPress plugin through version 9.14 allows unauthenticated remote attackers to trigger outbound HTTP requests to arbitrary URLs from the server, with third-party integration credentials automatically appended to each request. The vulnerability is conditional on the data-sync feature being enabled. A publicly available exploit exists per WPScan reporting, and successful exploitation results in full disclosure of stored integration credentials, making the confidentiality impact high despite a moderate aggregate CVSS score.
Improper access control in OSSRS SRS (Simple Realtime Server) before v5.0.213 allows remote attackers to bypass RTMP publish authorization enforced by the vhost security layer (security.enabled / SrsSecurity::check() in trunk/src/app/srs_app_security.cpp), enabling unauthorized stream publishing and, per the report, arbitrary code execution against the RTMP listener. The issue is tagged as both RCE and Authentication Bypass and affects deployments that rely on SRS's built-in allow/deny security rules to gate who may publish. Publicly available exploit code exists (a dedicated CVE-2026-68004 GitHub repository), though the vulnerability is not listed in CISA KEV and EPSS probability is low (0.42%, 35th percentile).
Remote code execution in Halo 2.25.4, an open-source Java-based CMS/blogging platform, arises from its plugin-installation machinery: the installFromUri method in PluginEndpoint.java combined with DefaultPluginApplicationContextFactory loads and executes attacker-supplied plugin code, giving full CWE-94 code injection with total confidentiality, integrity, and availability impact (CVSS 9.8). Publicly available exploit code exists in a referenced GitHub gist, but there is no public exploit identified as being actively used in the wild (not on CISA KEV). Note a signal conflict: the NVD vector states PR:N (unauthenticated) while plugin installation in Halo is normally an administrative function - this warrants verification.
Remote code execution in Systerel S2OPC 1.7.3 arises from a heap out-of-bounds write in the server-side EventFilter handling path exercised during CreateMonitoredItems processing. A remote attacker who can reach the OPC UA server endpoint and open a session on an events-enabled server can corrupt heap memory to achieve arbitrary code execution with the privileges of the server process. No public exploit identified at time of analysis, and EPSS is low (0.27%, 20th percentile), indicating the CVSS 9.8 rating reflects potential impact rather than observed mass exploitation.
Remote code execution in opensagres xdocreport (versions 0.9.2 through 2.2.0) arises because the bundled Velocity template engine ships with the insecure default UberspectImpl introspector, letting a crafted Velocity expression reach reflective JDK classes and run arbitrary code. Any application that renders untrusted template content through xdocreport's Velocity engine is exposed to full server compromise. Publicly available exploit code exists (a dedicated PoC GitHub repository and two HackMD write-ups), though the flaw is not listed in CISA KEV and no EPSS score was supplied.
Remote code execution in RainyGao DocSys 2.02.80 arises from an unrestricted file-upload handler (CWE-434) that fails to validate uploaded file type or extension, letting a remote attacker plant a server-executable payload (e.g. a webshell) and run arbitrary commands. The flaw carries a CVSS 9.8 with a fully remote, low-complexity, no-privilege, no-interaction vector, and a public technical write-up/POC exists in a third-party GitHub repository. It is not listed in CISA KEV, and EPSS is low at 0.26% (18th percentile), indicating no evidence of widespread automated exploitation yet.
Server-side request forgery in DataHub v1.5.0.1 lets a remote attacker coerce the server into fetching an attacker-supplied URL through its image-retrieval feature, with the response body and errors suppressed (blind SSRF). The reporter escalates this to arbitrary code execution, and the SSVC framework marks a proof-of-concept as available and the attack as automatable, though the flaw is not listed in CISA KEV and the EPSS probability is low at 0.24%.
Remote code execution in Squirro Cognitive Search before version 3.14.2 lets remote attackers run arbitrary code by sending a crafted payload to the application's password-reset function, a CWE-94 code-injection flaw. The password-reset endpoint is typically pre-authentication and internet-facing, and the CVSS vector (PR:N/UI:N) indicates no credentials or user interaction are needed. Publicly available exploit code exists (SSVC exploitation status 'poc', plus a public GitHub research repo), but this CVE is not in CISA KEV and EPSS is low at 0.24% (15th percentile), indicating credible but not yet widespread exploitation.
Memory-corruption remote code execution affects SteelSeries GG for macOS v107.0.0, where a classic buffer overflow (CWE-120) in the bundled libSSEdevice.dylib-specifically the CxAudioHidDevice::DeviceGetDescriptionString routine-lets an attacker overwrite memory while the client parses a device description string and thereby run arbitrary code in the context of the GG process. NVD rates it CVSS 9.8 with an unauthenticated network vector, though the affected component is a HID (USB audio device) descriptor parser, which points toward a malicious-device attack surface rather than pure network reachability. A third-party research writeup (Cipher-Security-Labs CSL-2026-03) is referenced, but there is no CISA KEV listing and EPSS is low at 0.21% (12th percentile).
Arbitrary code execution in SteelSeries GG for macOS v107.0.0 stems from a buffer overflow in the bundled libSSEdevice.dylib, specifically its dup_wcs (wide-character string duplication) routine, letting an attacker who can feed crafted input to the client run code with the privileges of the SteelSeries process. The flaw is disclosed by MITRE with an accompanying Cipher Security Labs research write-up (publicly available research/exploit material), but it is not on the CISA KEV list and EPSS is low at 0.21% (12th percentile). Note a material discrepancy: the CVSS vector is scored AV:N (network) yet the affected component is a local desktop peripheral-management library, which typically implies attacker-supplied data rather than an exposed network service.
OS command injection in the Tenda W20E router firmware version 16.01.0.6(2782) allows remote attackers to execute arbitrary code by abusing the url_need_login function, which fails to sanitize attacker-controlled input before passing it to a system shell (CWE-78). The flaw carries a CVSS 9.8 (AV:N/AC:L/PR:N/UI:N) rating, meaning unauthenticated network-based exploitation with full compromise of the device. Publicly available exploit code exists via a GitHub disclosure repository, though EPSS remains low (0.21%) and it is not listed in CISA KEV.
Remote code execution in JeecgBoot v3.9.2 allows attackers to run arbitrary code by supplying a malicious value to the 'files' parameter within the AI Chat module. The flaw is a code-injection weakness (CWE-94) carrying a critical 9.8 CVSS rating with a fully remote, low-complexity vector. There is no public exploit identified at time of analysis, and the EPSS probability is low (0.21%, 11th percentile), indicating no observed mass-exploitation activity yet.
Privilege escalation in GAPTEQ Designer v3.5 allows a remote attacker to abuse the Company Manager role to gain elevated privileges, per CWE-269 (improper privilege management). Publicly available exploit code exists (referenced on GitHub), and CISA's SSVC framework rates the technical impact as total and the flaw as automatable, though it is not listed in CISA KEV and EPSS remains low (0.21%, 11th percentile). The scope of affected deployments is limited to the GAPTEQ low-code application platform rather than a widely embedded library.
Privilege escalation in Squirro Cognitive Search before v3.14.2 lets a remote attacker submit a crafted request to gain elevated privileges within the enterprise search platform. The flaw is rated CVSS 9.8 with full confidentiality, integrity, and availability impact, and CISA's SSVC framework marks it automatable with total technical impact. No public exploit identified at time of analysis, and EPSS estimates near-term exploitation probability at just 0.21% (11th percentile).
SQL injection in Qcms 6.0.6 lets remote attackers inject arbitrary SQL through the backend tableField action, enabling database compromise and, per the advisory tagging, escalation toward arbitrary code execution. The flaw is documented in a white-box audit write-up but is not listed in CISA KEV, and its EPSS probability is low (0.20%, 10th percentile), indicating no evidence of widespread automated exploitation yet. Qcms is a niche PHP content-management system, so real-world exposure is limited despite the high CVSS score.
Unauthenticated remote root access affects the Tenda W20E router (firmware V16.01.0.6(2782)) through its /goform/telnet web endpoint, which activates the device's Telnet daemon without any authentication check. Any attacker who can reach the router's management interface can enable Telnet and log in for a root shell, giving full control of the device. A public disclosure/PoC repository exists, though the flaw is not listed in CISA KEV and EPSS-modeled exploitation probability is currently low (0.19%).
Remote code execution in the Tenda W20E router firmware V16.01.0.6(2782) lets unauthenticated network attackers overflow a stack buffer and run arbitrary code on the device. The flaw is explicitly an incomplete fix for the earlier CVE-2025-44867 and CVE-2026-36819, meaning the same code path remains exploitable despite prior patch attempts. A public vulnerability-disclosure repository exists, but the issue is not in CISA KEV and carries a low EPSS score (0.17%, 7th percentile), indicating no confirmed widespread exploitation yet.
SQL injection in Brainformatik CRM+ through version 2025.6 allows remote unauthenticated attackers to inject arbitrary SQL via the appointment conflict-checking AJAX endpoint (index.php?module=Appointments&action=CheckConflictOfDates), enabling full database compromise. The flaw is time-based (blind) but scores CVSS 9.8 with total technical impact; a public proof-of-concept exists (CISA SSVC: exploitation=poc) though it is not yet in CISA KEV. EPSS is low (0.16%, 6th percentile), indicating no widespread automated exploitation observed to date despite the critical severity.
Arbitrary code execution is reported in PbootCMS 3.2.15, a PHP-based open-source content management system, reachable through several front-end controllers (MemberController.php, UserController.php, CommentController.php, ContentController.php) and the shared helper.php. Classified as CWE-94 code injection with a CVSS 9.8 and tagged PHP/RCE, it would let an attacker run attacker-supplied PHP on the web server. No public exploit is identified at time of analysis and it is not listed in CISA KEV; note that one of the two primary references is a document titled 'URL Redirection Vulnerability,' which conflicts with the code-execution description and should be verified.
Unauthorized internal-account access in Mahara ePortfolio (all releases before 25.04.5 and 26.04.0) allows remote attackers to log into other users' accounts by abusing the Learning Tools Interoperability (LTI) integration, an authentication-bypass class flaw affecting both the LTI 1.1 and LTI 1.3 Advantage connectors. The issue occurs 'under certain circumstances,' implying LTI must be configured and reachable, and no public exploit identified at time of analysis. The assigned CVSS of 9.8 reflects full confidentiality/integrity/availability compromise of the targeted account, though the conditional wording suggests real-world exploitation depends on LTI being enabled.
Information disclosure in the Mahara ePortfolio platform before 25.04.5 and 26.04.0 allows remote attackers to retrieve backed-up content belonging to a different Text block/section by crafting a request in a specific way. The flaw (CWE-200) exposes stored draft or historical text that should be isolated per section, potentially leaking other users' or other pages' content. No public exploit has been identified at time of analysis, and the vulnerability is not listed in CISA KEV.
Denial-of-service in TOTOLINK EX1200L firmware 9.3.5u.6146_B20201023 allows authenticated remote attackers to crash the device by sending a crafted request to the setWizardCfg function in /cgi-bin/cstecgi.cgi, triggering a null pointer dereference. Impact is limited strictly to availability - no confidentiality or integrity compromise - but a publicly available proof-of-concept exploit on GitHub (CVSS 4.0 E:P) significantly lowers the exploitation barrier. No vendor patch has been identified and the device is not listed in CISA KEV, making access restriction to the management interface the primary mitigation.
Remote denial-of-service in TOTOLINK EX1200L firmware 9.3.5u.6146_B20201023 allows a low-privileged authenticated attacker to crash the device by submitting a malformed request to the setPasswordCfg function in the CGI-based web management interface. Exploitation triggers a null pointer dereference in /cgi-bin/cstecgi.cgi, rendering the device unresponsive and disrupting its network extension services entirely. A public proof-of-concept exploit exists on GitHub, lowering the bar for exploitation, though no active exploitation has been recorded in CISA KEV at time of analysis.
Cross-user credential disclosure in Onyx (open-source AI platform) lets any low-privileged BASIC_ACCESS user retrieve another user's OAuth Authorization header via the GET /api/mcp/servers and GET /api/mcp/servers/persona/{persona_id} MCP endpoints. Per-user OAuth tokens were incorrectly persisted into a shared admin MCPConnectionConfig row and then echoed back through auth_template.headers to unrelated callers, leaking bearer tokens for downstream MCP services. Affects versions prior to 3.1.10, 3.2.14, and 4.0.0; no public exploit identified at time of analysis and it is not listed in CISA KEV, but a vendor patch and detailed fix commits are available.
SQL injection in SourceCodester Pet Grooming Management Software 1.0 exposes the /admin/get_barcode_data.php endpoint to remote attackers who can manipulate the unsanitized 'barcode' parameter to read, alter, or corrupt the underlying database. The CVSS 4.0 vector assigns no authentication or user-interaction requirement, and a publicly available proof-of-concept lowers the exploitation bar to script-level effort. No CISA KEV listing exists, but the combination of a network-exposed attack surface and published exploit code makes this a higher practical priority than the nominal Medium score implies for any internet-facing deployment.
SQL injection in SourceCodester Class and Exam Timetabling System 1.0 exposes the `/edit_subject1.php` endpoint to unauthenticated remote attackers who can manipulate the `ID` parameter to execute arbitrary SQL against the underlying database. The vulnerability carries a CVSS 4.0 score of 6.9 with no authentication or user interaction required, and a proof-of-concept exploit has been publicly disclosed via GitHub. No KEV listing exists, but the public POC materially lowers the barrier to exploitation for opportunistic attackers targeting exposed instances.
SQL injection in SourceCodester Class and Exam Timetabling System 1.0 exposes the `/edit_subject2.php` endpoint to remote unauthenticated attackers who can manipulate the `ID` parameter to execute arbitrary SQL commands against the backend database. The CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N) confirms exploitation requires no authentication, no user interaction, and no special attack prerequisites. Public exploit code exists per the referenced GitHub issue, lowering the skill bar for attackers, though no CISA KEV listing confirms widespread active exploitation at time of analysis.
Stored cross-site scripting in SiYuan before v3.7.4 lets an authenticated user upload an HTML file through the asset-serving endpoint, which the server returns without Content-Disposition or X-Content-Type-Options headers, so the browser renders it inline instead of downloading it. When the workspace owner clicks the malicious asset link, the injected script runs in the app's origin with full kernel API access, enabling account and data compromise. Reported by VulnCheck with a vendor patch available; no public exploit has been identified at time of analysis.