Remote unauthenticated attackers achieve full code execution on Paperclip AI orchestration servers (versions prior to 2026.416.0) via authentication bypass through a six-step API call chain. The attack requires no credentials, no user interaction, and succeeds against default 'authenticated' mode deployments exposed to network access. CVSS 10.0 with scope change indicates container/host escape potential. No active exploitation confirmed in CISA KEV at time of analysis, though the vendor advisory (GitHub Security Advisory GHSA-68qg-g8mg-6pr7) confirms the critical authentication bypass mechanism in both @paperclipai/server and paperclip npm packages.
Arbitrary file upload in Breeze Cache for WordPress allows unauthenticated remote attackers to upload malicious files and achieve remote code execution on vulnerable servers. Exploitation requires the non-default 'Host Files Locally - Gravatars' feature to be enabled. While CVSS rates this 9.8 critical, real-world exposure is limited by the disabled-by-default configuration requirement. No public exploit code or CISA KEV listing identified at time of analysis, though Wordfence threat intelligence has disclosed technical details including vulnerable code paths.
OS command injection in radare2-mcp (the Model Context Protocol server that bridges the radare2 reverse-engineering framework to LLM agents) versions 1.6.0 and earlier lets remote, unauthenticated attackers achieve full remote code execution on the host. Attacker-controlled JSON-RPC parameters flow into r2_cmd_str() where shell metacharacters bypass the command filter, and because the fix re-enables radare2's exec/socket sandbox, the flaw effectively granted attackers direct shell access. Publicly available exploit code exists (reported by VulnCheck, POC referenced in issue #45); it is not listed in CISA KEV and EPSS is modest at 0.27% (51st percentile).
Unauthenticated remote file read/write and NTLM coercion in Kofax Capture (rebranded Tungsten Capture) 6.0.0.0 stems from the Ascent Capture Service exposing a deprecated .NET Remoting HTTP channel on TCP/2424 with a default, publicly known endpoint identifier and no authentication. A remote attacker can abuse .NET Remoting object unmarshalling to instantiate a server-side System.Net.WebClient, allowing arbitrary file read/write and coerced NTLMv2 authentication that can cascade into credential theft, denial of service, remote code execution, or lateral movement depending on the service account's privileges. Publicly available exploit code exists (VulnCheck/gist PoC); it is not listed in CISA KEV, and EPSS is low at 0.16% (37th percentile).
Remote code execution in KTransformers through 0.5.3 lets unauthenticated network attackers run arbitrary code by sending a crafted pickle payload to the scheduler RPC service used in balance_serve backend mode. The affected code binds a ZMQ ROUTER socket to all interfaces (0.0.0.0) with no authentication and passes received frames straight into pickle.loads(), so any host that can reach the scheduler port gains code execution as the ktransformers process. Publicly available exploit code exists (VulnCheck/chocapikk write-up) and a vendor upstream fix is available, though the EPSS probability is currently low at 0.04%.
Local privilege escalation in Tenable Nessus (≤10.11.3) and Nessus Agent (≤11.1.2) on Windows lets a low-privileged user abuse an NTFS junction to force the SYSTEM-level Nessus service into deleting arbitrary files, which can be leveraged for code execution as SYSTEM. Publicly available exploit code exists, but there is no evidence of active exploitation and the EPSS probability is negligible (0.01%, 3rd percentile). SSVC rates technical impact as total but exploitation as none and non-automatable, making this a credible but not urgent operator-side EoP.
Authentication bypass via path traversal in Intrado 911 Emergency Gateway (EGW) versions 5.x, 6.x, and 7.x allows an attacker with existing network access to reach the EGW management interface without valid credentials and read, modify, or delete files. There is no public exploit identified at time of analysis and it is not listed in CISA KEV, but CISA's SSVC assessment rates the technical impact as total and the flaw as automatable, and the CVSS 4.0 base score is 9.3 (critical). EPSS probability is low (0.17%, 38th percentile), indicating no observed mass-exploitation activity yet.
Client-side authentication bypass in SenseLive X3050's web management interface allows remote unauthenticated attackers to gain full administrative access by extracting hardcoded credentials from browser-executed JavaScript. The vulnerability enables complete compromise of device management with zero technical barriers (CVSS 9.3, AV:N/AC:L/PR:N). CISA ICS-CERT has published an advisory, indicating this affects operational technology environments where administrative access to industrial sensors could enable process manipulation or monitoring disruption.
Authentication bypass in SenseLive X3050 web management interface allows remote unauthenticated attackers to gain administrative access using default or previously-set credentials. After factory restore via SenseLive Config 2.0 tool, password updates fail to propagate correctly - the interface falsely reports success while the backend continues accepting old credentials. CISA ICS-CERT has issued an advisory (ICSA-26-111-12), indicating this affects industrial control system deployments. With CVSS 9.3 (AV:N/AC:L/PR:N) and CWE-522 (Insufficiently Protected Credentials), this represents critical risk for remotely accessible devices where administrators believe credentials have been changed but remain exploitable.
Remote unauthenticated attackers can permanently disable SenseLive X3050 industrial gateways and connected RS-485 downstream systems by modifying critical configuration parameters through the web management interface. The device's lack of physical reset button forces specialized console-based factory reset procedures, making this a high-impact operational disruption vector for industrial environments. CISA ICS-CERT has issued an advisory (ICSA-26-111-12), indicating industrial sector awareness of this authentication bypass flaw.
Path traversal in radare2's project deletion function (r_core_project_delete in libr/core/project.c) allows a local attacker to cause recursive deletion of arbitrary directories by supplying crafted absolute project paths that escape the configured dir.projects root. All versions prior to 6.1.4 are affected. An attacker who controls a project file path can cause radare2 - running with the victim's permissions - to wipe arbitrary directories, resulting in significant integrity and availability impact. No public exploit identified at time of analysis for weaponized in-the-wild use; however, a publicly available exploit (POC) exists via the upstream GitHub PR, and SSVC rates technical impact as total despite non-automatable exploitation.
Path traversal in radare2's project notes handling exposes arbitrary filesystem files to read and overwrite when a user imports a crafted .zrp archive. All radare2 releases prior to 6.1.4 are vulnerable via the `r_core_project_notes_file()` function in `libr/core/project.c`, which failed to detect symlinks within imported project archives before performing note operations. An attacker who can deliver and persuade a victim to import a malicious .zrp file can escape the configured `dir.projects` directory entirely, gaining low confidentiality and high integrity impact on the host filesystem. No active exploitation is confirmed in CISA KEV, but a working proof-of-concept is publicly available in the referenced GitHub PR test suite.
Broken access control in SpiceJet's Online Booking System exposes complete passenger records through the public booking-retrieval page, which requires only a PNR code and passenger last name with no authentication or ownership verification. Any unauthenticated internet user who obtains or guesses these two inputs can retrieve extensive personal, travel, and booking metadata. No public exploit has been identified at time of analysis and EPSS is low (0.06%), but CISA's SSVC framework rates the flaw as automatable, meaning enumeration at scale is technically feasible.
Unauthorized passenger data disclosure in SpiceJet's Online Booking System lets unauthenticated remote attackers retrieve passenger name records (PNRs) and associated names by directly querying a profile endpoint that lacks authorization checks. Because PNR identifiers are predictable, records can be enumerated in bulk to harvest passenger names at scale. There is no public exploit identified at time of analysis and EPSS is very low (0.05%), but CISA's SSVC assessment flags the flaw as automatable with partial technical impact, and it is tracked as an ICS advisory (ICSA-26-113-04) and EU EUVD-2026-25299.
Remote code execution in Microsoft Bing lets an unauthenticated network attacker run arbitrary code by submitting untrusted serialized data that the service deserializes unsafely (CWE-502). The flaw carries a maximum CVSS 3.1 base score of 10.0 with a scope change, meaning exploitation of the Bing service could impact resources beyond its own security authority. There is no public exploit identified at time of analysis and it is not in CISA KEV; EPSS is a modest 0.27% (51st percentile), reflecting probability rather than confirmed activity.
Server-side request forgery in Microsoft Entra ID Entitlement Management lets a remote, unauthenticated attacker coerce the cloud service into making attacker-directed requests, enabling spoofing and full compromise of confidentiality, integrity, and availability across a security-scope boundary (CVSS 10.0). No public exploit identified at time of analysis and EPSS is low (0.07%), but the SSVC framework flags the flaw as automatable with total technical impact. As a cloud-hosted Microsoft identity service, remediation is delivered by the vendor rather than by customer patching.
Authenticated customers can achieve remote code execution in Froxlor server administration software versions prior to 2.3.6 through path traversal in the API's language parameter. By injecting malicious path traversal sequences into the `def_language` field via the `Customers.update` or `Admins.update` API endpoints, authenticated users can force the application to execute arbitrary PHP code as the web server user on subsequent requests. This vulnerability carries a CVSS score of 9.9 with scope change, indicating potential for full system compromise beyond the vulnerable component. Vendor-released patch version 2.3.6 addresses the vulnerability by implementing proper validation of language parameters against available language files.
Stored cross-site scripting in hackage-server (the Haskell Hackage package registry software) lets a low-privileged authenticated package uploader inject persistent JavaScript by placing crafted metadata in a package's .cabal file, which the server renders into HTML href attributes without sanitization (CWE-79). Because the payload persists in package pages, any user or maintainer viewing the affected package - potentially including administrators - can have script executed in their session. No public exploit has been identified at time of analysis and EPSS is very low (0.05%, 14th percentile), but SSVC rates the technical impact as total.
Stored cross-site scripting in hackage-server (the software powering hackage.haskell.org) allowed HTML and JavaScript embedded in source packages or documentation uploads to be served unsanitized from the primary hackage.haskell.org origin. A low-privileged, authenticated package maintainer could plant a malicious payload so that any authenticated user visiting the affected package or documentation pages would have their active session hijacked, letting the attacker upload packages/docs, alter maintainer lists and metadata, or perform any action the victim is authorized to do. There is no public exploit identified at time of analysis, and EPSS is very low (0.05%), but the CVSS 9.9 rating reflects the total technical impact of a same-origin XSS on a central package-signing authority.
Remote code execution in FunnelFormsPro WordPress plugin (versions up to 3.8.1) allows authenticated attackers to inject and execute arbitrary code on vulnerable servers. The CVSS 9.9 Critical rating reflects the scope change (S:C) and complete system compromise (C:H/I:H/A:H). Exploitation requires low-privilege authentication (PR:L) but no user interaction, making it exploitable by subscriber-level WordPress accounts. EPSS and KEV status not provided in available data, limiting real-world exploitation confidence assessment.
Remote code execution in the Pipecat voice/multimodal agent framework (Python) affects versions 0.0.41 through 0.0.93 when the optional, non-default LivekitFrameSerializer is enabled, because its deserialize() method feeds raw WebSocket bytes into pickle.loads() without validation. A malicious WebSocket client can send a crafted pickle payload to run arbitrary code on the server, and if the service listens on an external interface (e.g. 0.0.0.0) the attack is reachable over the network or internet. The issue is fixed in 0.0.94 (serializer removed); publicly available exploit code exists per the SSVC 'poc' status, but it is not listed in CISA KEV.
Remote code execution in H2O-3 versions 3.46.0.9 and earlier allows unauthenticated attackers to execute arbitrary code via the /99/ImportSQLTable REST API by abusing PostgreSQL JDBC driver parameters that bypass an incomplete MySQL-only parameter blacklist. No active exploitation is recorded in CISA KEV and EPSS is low (0.19%), but a vendor patch is available and SSVC marks exploitation status as POC, indicating proof-of-concept-grade attacker capability against a network-reachable endpoint.
Remote code execution in Flowise before 3.1.0 lets an unauthenticated attacker run arbitrary OS commands as root inside the container with a single HTTP request. The flaw chains a parameter-override validation bypass (the FILE-STORAGE:: keyword check) with NODE_OPTIONS environment-variable injection through the Custom MCP node. Publicly available exploit code exists per the SSVC 'poc' rating, though EPSS is only 0.17%; no active exploitation has been confirmed.
OS command injection in the TOTOLINK A3300R router (firmware v17.0.0cu.557_B20221024) lets remote attackers run arbitrary shell commands by injecting into the stunMinAlive parameter of the /cgi-bin/cstecgi.cgi endpoint. The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N) scores it 9.8 Critical with no authentication qualifier indicated, and CISA's SSVC framework marks it automatable with total technical impact. Publicly available exploit code exists (researcher writeup on GitHub), though EPSS remains very low at 0.06% and it is not on the CISA KEV list - no public exploit indicates active exploitation yet.
OS command injection in TOTOLINK A3300R routers (firmware v17.0.0cu.557_B20221024) lets remote attackers run arbitrary commands by injecting shell metacharacters into the stunServerAddr parameter of the /cgi-bin/cstecgi.cgi endpoint. The flaw carries CVSS 9.8 (AV:N/AC:L/PR:N/UI:N) and publicly available exploit code exists, though EPSS remains low at 0.06% and it is not listed in CISA KEV.
OS command injection in the TOTOLINK A3300R router firmware (v17.0.0cu.557_B20221024) lets remote attackers run arbitrary system commands by injecting shell metacharacters into the stunMaxAlive parameter handled by the /cgi-bin/cstecgi.cgi endpoint. The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N) describes unauthenticated network exploitation with total confidentiality, integrity, and availability impact, effectively granting full control of the device. Publicly available exploit code exists (GitHub PoC), though EPSS is very low (0.06%) and it is not listed in CISA KEV, so exploitation is proof-of-concept rather than confirmed in-the-wild.
Command injection in the TOTOLINK A3300R router (firmware v17.0.0cu.557_B20221024) lets remote attackers execute arbitrary OS commands by injecting into the stunEnable parameter of the /cgi-bin/cstecgi.cgi endpoint. The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N) indicates unauthenticated network exploitation with total confidentiality, integrity, and availability impact, and publicly available exploit code exists via a GitHub PoC. Despite the 9.8 rating, EPSS is only 0.06% (19th percentile), and there is no CISA KEV listing, so no active exploitation is confirmed.
Use-after-free in the Linux kernel's kernel TLS (net/tls) data path lets a local attacker corrupt kernel memory by racing the -EBUSY crypto-backlog error path of tls_do_encryption(). When crypto_aead_encrypt() returns -EBUSY, both the async tls_encrypt_done() callback and the synchronous error path clean up encrypt_pending and the scatterlist, double-decrementing a sentinel so a later sendmsg frees an in-flight tls_rec while a cryptd callback is still pending. Despite an NVD CVSS of 9.8, EPSS is 0.01% and there is no public exploit identified at time of analysis, reflecting the narrow race conditions required.
NoSQL injection in Rocket.Chat before 8.3.0 (and the 8.2.1, 8.1.2, 8.0.3, 7.13.5, 7.12.6, 7.11.6, and 7.10.9 maintenance releases) lets remote unauthenticated attackers take over the account of the first user holding a generated OAuth token when the built-in OAuth2 server has an app configured. The unsanitized access_token query parameter is passed into a MongoDB lookup, so operator injection (e.g. access_token[$ne]=null) matches a valid token without knowing it, authenticating the attacker as that user — typically the initial admin. There is no public exploit identified at time of analysis, though the fixing pull request and its regression tests openly disclose the exact injection payloads; EPSS is very low (0.01%, 3rd percentile) and CISA SSVC records no observed exploitation.
SQL injection in JizhiCMS v2.5.4 lets attackers inject arbitrary SQL through the product editing module, exploiting unsanitized input handled by CWE-89. The flaw carries a 9.8 CVSS with total technical impact (SSVC), and publicly available exploit code exists via a shared gist, though EPSS remains very low (0.01%) and it is not on CISA KEV. Because the product editing module is a backend/administrative function, real-world abuse likely presupposes access to that interface despite the network-facing CVSS scoring.
Sandbox escape in Google Chrome's GPU process on Android (versions prior to 147.0.7727.117) can be triggered by an attacker who has already compromised the renderer process, using a crafted HTML page to force an out-of-bounds read (CWE-125) that leaks GPU-process memory and potentially breaks out of the renderer sandbox. Rated High by Chromium and CVSS 9.6, it is a second-stage bug requiring prior renderer compromise. No public exploit identified at time of analysis, and EPSS is low at 0.07% (21st percentile).
Privilege escalation in Microsoft Partner Center allows an authenticated, low-privileged user to elevate their access over the network due to improper access control (CWE-284), crossing a security boundary (CVSS scope-changed, 9.6). With confidentiality and integrity impact rated High and a scope change, a successful attacker can reach data or functions belonging to other tenants or higher privilege levels within the Partner Center platform. There is no public exploit identified at time of analysis, EPSS is very low (0.04%), and CISA's SSVC lists exploitation status as none.
Sandbox escape in Google Chrome's DevTools component (versions prior to 147.0.7727.117) lets an attacker who has already compromised the renderer process break out of the browser sandbox via a crafted HTML page. Rated High by Chromium and CVSS 9.6, this is a use-after-free (CWE-416) that serves as the second link in a multi-stage exploit chain rather than a standalone remote code execution. No public exploit identified at time of analysis, EPSS is very low (0.03%), and CISA SSVC records no observed exploitation, so it is a patch-and-move-on desktop browser fix rather than an emergency.
Cross-Site Request Forgery in hackage-server (the Haskell Hackage package repository server) allows a malicious web page to force a victim's authenticated browser to issue state-changing requests, letting attackers upload packages or perform administrative actions using the victim's latent session, and to abuse unauthenticated flows such as creating new user accounts. The flaw stems from a site-wide absence of CSRF tokens across endpoints. There is no public exploit identified at time of analysis, and EPSS exploitation probability is very low (0.02%, 4th percentile), but technical impact is rated total.
Authenticated remote code execution in FlowiseAI Flowise (and the flowise-components package) prior to 3.1.0 arises from the CSVAgent node, which interpolates a user-supplied 'customReadCSV' Pandas string directly into Python code executed via pyodide. An authenticated attacker (or an unauthenticated one when FLOWISE_USERNAME/FLOWISE_PASSWORD are unset, via the 'x-request-from: internal' header bypass) can inject Python such as 'import os;os.system(...)' to run arbitrary commands on the server. Publicly available exploit code exists (PoC in the GHSA advisory); this is not listed in CISA KEV, and EPSS is modest at 0.62% (70th percentile).
HT Mega Addons for Elementor WordPress plugin before version 3.0.7 leaks customer PII through an unauthenticated AJAX endpoint, exposing full name, city, state, and country for any orders placed in the prior 7 days. Any remote unauthenticated attacker can query this endpoint without credentials on affected WordPress installations. A public proof-of-concept is available and SSVC classifies the attack as automatable, enabling mass scanning of exposed WooCommerce stores; however, the EPSS score of 0.77% suggests exploitation has not yet become widespread.
Remote code execution in Borg SPM 2007 allows unauthenticated attackers to upload and execute web shell backdoors via unrestricted file upload vulnerability. This discontinued product (sales ended 2008) remains exploitable over the network with no authentication required, enabling full server compromise. CVSS 9.3 (Critical) with network vector, low complexity, and no privileges required. EPSS and KEV data not available for this CVE, but the trivial attack requirements (AV:N/AC:L/PR:N/UI:N) indicate high exploitability if exposed systems exist.
Unauthenticated SQL injection in SocialEngine (versions ≤7.8.0) via the text parameter of the /activity/index/get-memberall endpoint lets remote attackers read arbitrary database contents, reset administrator passwords, and reach the Admin Panel's Packages Manager - a path that can escalate to remote code execution. A public proof-of-concept exists (KarmaInSecurity KIS-2026-08) and SSVC rates the flaw as automatable with total technical impact, though EPSS remains low at 0.17%. No public exploit in CISA KEV, so this is publicly available exploit code rather than confirmed active exploitation.
Authentication bypass in Borg SPM 2007 allows remote unauthenticated attackers to impersonate any user and gain complete system access without credentials. This discontinued product (sales ended 2008) presents maximum network exposure (CVSS:4.0 9.3, AV:N/AC:L/PR:N) with trivial exploitation conditions. While no CISA KEV listing exists, the simplicity of exploitation combined with complete system compromise (VC:H/VI:H/VA:H) makes this critical for organizations still running this legacy software, though real-world deployment is likely minimal given the 18-year product discontinuation.
SQL injection in Borg SPM 2007 allows unauthenticated remote attackers to execute arbitrary SQL commands via network requests, enabling complete database compromise including read, modify, and delete operations. This legacy product (sales ended 2008) receives a critical CVSS 9.3 score with network vector, low complexity, and no authentication required. Taiwan CERT issued advisories identifying this as a SQL injection vulnerability affecting an end-of-life business management system, though no active exploitation evidence (KEV) or public exploit code has been identified at time of analysis.
Cypher injection in Flowise GraphCypherQAChain node allows remote unauthenticated attackers to execute arbitrary database commands against connected Neo4j instances. Attackers can exfiltrate, modify, or delete data in the graph database by injecting malicious Cypher queries through user-controlled input fields that bypass sanitization (CWE-943: Improper Neutralization of Special Elements in Data Query Logic). The vulnerability affects both Flowise core and flowise-components packages prior to version 3.1.0. CVSS 9.3 critical severity reflects network-accessible attack vector requiring no authentication or user interaction. EPSS data unavailable; no CISA KEV listing indicates exploitation not yet confirmed in the wild, though GitHub security advisory confirms vendor awareness and patch availability.
Unauthenticated remote code execution in HuggingFace LeRobot (versions 0 through 0.5.1) stems from pickle.loads() being called on attacker-controlled bytes received over the async inference gRPC pipeline, which runs without authentication or TLS. Any network-reachable attacker can execute arbitrary code on either the policy server or the robot client by delivering a malicious pickle payload through the SendPolicyInstructions, SendObservations, or GetActions RPCs. Publicly available exploit code exists (chocapikk write-up) and SSVC rates technical impact as total and the flaw as automatable, though EPSS remains very low (0.06%) reflecting limited real-world telemetry rather than low technical severity.
Privilege elevation in Microsoft 365 Copilot via an open redirect (CWE-601) lets an unauthenticated remote attacker abuse the AI assistant's URL-handling to bounce a victim to an attacker-controlled site and elevate privileges over the network. Rated CVSS 9.3 with a scope-changed vector because the redirect crosses a trust boundary, though exploitation requires the victim to interact (click) with the crafted link. No public exploit identified at time of analysis and EPSS is very low (0.04%), indicating no observed real-world exploitation despite the high severity score.
Server-side request forgery in Microsoft Dynamics 365 (Online) lets a remote, unauthenticated attacker coerce the cloud service into issuing crafted requests to perform network spoofing, with the CVSS 3.1 score of 9.3 driven by a scope change (S:C) and high confidentiality and integrity impact. Because Dynamics 365 (Online) is a Microsoft-hosted SaaS, exploitation requires victim user interaction (UI:R) and the fix is applied server-side by the vendor. There is no public exploit identified at time of analysis, EPSS is very low (0.04%, 13th percentile), and CISA SSVC scores exploitation as 'none', indicating no observed in-the-wild activity despite the high base score.
Replay-based keyless entry bypass in Yadea T5 electric bicycles (2024 and later model years) lets a nearby attacker clone the RF signal from the key fob and take full control of the vehicle. Because the fob uses the EV1527 fixed-code protocol with no rolling codes or challenge-response, a single captured transmission can be rebroadcast to unlock and operate the bike. Publicly available exploit code exists (two GitHub repositories), though EPSS is very low (0.03%) and it is not listed in CISA KEV.
Remote code execution in FlowiseAI Flowise before 3.1.0 lets an unauthenticated attacker run arbitrary OS commands on the server via the Airtable Agent node. Because LLM-generated Python is executed without proper sandboxing, an attacker who can send prompts to a chatflow containing this node can use prompt injection to coerce the model into emitting malicious Python that runs in the context of the Flowise process. Reported through Trend Micro ZDI (ZDI-CAN-29412); publicly available exploit code exists (SSVC: PoC), though EPSS is low at 0.06% and it is not listed in CISA KEV.
Unauthorized configuration tampering in SenseLive X3050 web management interface allows authenticated attackers to set critical system parameters (IP addressing, watchdog timers, reconnect intervals, service ports) to unsafe values, causing persistent device unavailability or operational instability. CISA ICS-CERT advisory confirms impact on industrial control systems. Network-accessible with low complexity (AV:N/AC:L) but requires low-privilege authentication (PR:L). High integrity and availability impact (VI:H/VA:H) with zero confidentiality impact. No public exploit identified at time of analysis.
Remote code execution in Froxlor server administration software versions prior to 2.3.6 allows authenticated administrators with change_serversettings permission to inject arbitrary PHP code through an unescaped MySQL server configuration parameter. The vulnerability enables persistent code execution on every subsequent HTTP request as the web server user due to improper input sanitization in PhpHelper::parseArrayToString(). Vendor patch available in version 2.3.6. CVSS score of 9.1 reflects the critical impact despite requiring high-privilege authentication, with scope change indicating the attacker can break out of the application's security context.
Remote code execution in Luanti 5.0.0 through 5.15.1 allows authenticated attackers to escape the Lua sandbox via malicious mods, achieving arbitrary code execution and full filesystem access on victim devices when LuaJIT is enabled. The vulnerability affects server-side mods, async/mapgen environments, and client-side mods (CSM), requiring only low privileges to exploit. A vendor patch is available in version 5.15.2, addressing a CWE-94 code injection flaw that enables complete compromise of the host system. No active exploitation or proof-of-concept has been publicly identified at time of analysis.
Command injection in elFinder before 2.1.67 lets attackers reach OS command execution as the web server user through the resize command's bg (background color) parameter. When the file manager is configured with the ImageMagick CLI backend, the user-supplied bg value is interpolated into shell command strings without sufficient escaping, allowing shell metacharacter injection. Fixed in 2.1.67; no public exploit is identified at time of analysis, though the CVSS 4.0 threat metric E:P indicates proof-of-concept maturity, EPSS is low (0.44%), and it is not in CISA KEV.