Remote authenticated attackers can execute arbitrary code on Tenda HG10 routers (firmware HG7_HG9_HG10re_300001138_en_xpon) by sending a malformed 'nextHop' parameter to the /boaform/formRouting endpoint in the Boa web service. This buffer overflow vulnerability has publicly available exploit code on GitHub and is rated 8.8 (High) with low attack complexity. EPSS data unavailable; not currently listed in CISA KEV. Successful exploitation requires only low-privilege authentication and grants full device compromise (confidentiality, integrity, and availability impact).
Remote command injection in GitPilot-MCP allows unauthenticated attackers to execute arbitrary system commands via the repo_path function in main.py. The vulnerability affects all versions up to commit 9ed9f153ba4158a2ad230ee4871b25130da29ffd, with publicly available exploit code demonstrating practical exploitation. CVSS 7.3 (High) with network vector and no authentication required indicates significant exposure, though CVSS impact ratings (L/L/L) suggest attackers may have limited privileges in command execution context.
Remote denial of service in Cesanta Mongoose up to version 7.20 allows unauthenticated attackers to trigger an infinite loop via manipulation of TCP option length parameters in the handle_opt function, causing service unavailability. Publicly available exploit code exists. Patch released in version 7.21.
Improper authorization in Vanna AI's Legacy Flask API allows remote unauthenticated attackers to bypass authentication controls and gain unauthorized access. Affects Vanna versions up to 2.0.2. A publicly available proof-of-concept exploit has been disclosed on GitHub, though vendor has not responded to disclosure. CVSS 7.3 severity reflects network-accessible attack with low complexity requiring no privileges or user interaction, enabling confidentiality, integrity, and availability impacts.
Unintended intermediary exposure in go-kratos kratos up to 2.9.2 allows remote attackers to disclose sensitive information via manipulation of the http.DefaultServeMux fallback handler in the NewServer function. The vulnerability has publicly available exploit code and affects the HTTP transport layer with a CVSS score of 5.5, representing a confidentiality impact without availability or integrity concerns.
Remote authentication bypass in note-mark backend allows unauthenticated attackers to hijack OIDC user accounts by submitting the password 'null' to the internal login endpoint. Affected deployments running default configuration (EnableInternalLogin=true) with OIDC enabled permit complete account takeover of any OIDC-registered user. Attackers gain full access to private notebooks, markdown content, and uploaded assets, plus can persist access by overwriting the victim's password. Vendor patch available in commit dea5530c. CVSS 9.4 (AV:N/AC:L/PR:N/UI:N) reflects the zero-interaction remote attack against default installations. No EPSS or KEV data available, but the detailed POC script in the advisory significantly lowers exploitation barrier.
Out-of-bounds memory access in the Linux kernel's netfilter ip6t_eui64 IPv6 match module allows a malformed Ethernet/MAC header to reach eth_hdr(skb) before validation, leading to information disclosure or a kernel crash. The prior guard in eui64_mt6() only rejected an invalid MAC header when par->fragoff != 0, so non-fragmented packets (fragoff == 0) with an invalid MAC header could still be dereferenced. Exploitation requires an administrator to have configured an ip6tables rule using the eui64 match; no public exploit identified at time of analysis and EPSS is very low (0.02%).
Out-of-bounds read in the Linux kernel's bridge neighbour-discovery proxy (br_nd_send) allows remote attackers to leak adjacent kernel memory or crash the host by sending crafted IPv6 Neighbor Solicitation packets whose ND option area lies in the non-linear (paged/fragmented) part of the socket buffer. The function parses ns->opt[] assuming options reside in the skb's linear region, but callers only guarantee the ICMPv6 header and target address are linear, so parsing reads past the linear buffer. EPSS is very low (0.03%, 8th percentile) and there is no public exploit identified at time of analysis, but the fix is shipping across all maintained stable trees and downstream vendors (Ubuntu USN-8490-1, Siemens).
Privilege escalation in bubblewrap 0.11.x when installed setuid root allows local attackers to escape sandbox isolation via ptrace attachment during low-privileged setup phases. The vendor confirmed active exploitation risk by releasing emergency v0.11.2 patch and immediately deprecating setuid mode entirely. CVSS 8.7 severity reflects high integrity impact from sandbox breakout, though exploitation requires local access to a setuid-configured bubblewrap binary (non-default in most distributions).
Remote code execution in simple-git before 3.36.0 allows unauthenticated attackers to execute arbitrary commands via incomplete sanitization of command-line options. The vulnerability bypasses the prior CVE-2022-25912 fix by accepting --config instead of the blocked -c flag, enabling protocol.ext.allow=always configuration and malicious ext:: URLs. Publicly available exploit code exists (POC confirmed), with EPSS score of 0.08% indicating low current exploitation probability despite the theoretical severity. SSVC framework classifies this as automatable with total technical impact.
WireGuard private keys leak through Cilium debugging tools in deployments using transparent encryption. Cilium's cilium-bugtool and cilium sysdump command expose the node-to-node WireGuard encryption private key (cilium_wg0.key) in output archives. Attackers with high-privilege local access to these diagnostic outputs can decrypt past and future inter-node traffic. Affects all Cilium versions prior to v1.17.15, v1.18.0-v1.18.8, and v1.19.0-v1.19.2. Patches released in v1.17.15, v1.18.9, and v1.19.3. No public exploit identified at time of analysis, but exploitation requires only access to previously shared bugtool/sysdump archives.
Local privilege escalation and memory corruption in the Linux kernel's batman-adv mesh networking module allows a low-privileged attacker to trigger an skb_put buffer overflow by exploiting a race in OGM (Originator Message) aggregation. When aggregation is toggled at runtime, a forwarded packet allocated with only packet_len bytes can later be selected for aggregation, causing an append past the socket buffer's tailroom. CVSS 7.8 (High); EPSS is very low at 0.02% (7th percentile) and there is no public exploit identified at time of analysis and no CISA KEV listing.
Use-after-free in the Linux kernel IPv6 flowlabel subsystem lets a local user crash the system by racing a read of /proc/net/ip6_flowlabel against the early free of an exclusive flowlabel's option block. When fl_release() drops fl->users to zero it immediately kfree()s fl->opt, while the enclosing struct ip6_flowlabel stays visible in the global hash until later RCU garbage collection, so a concurrent ip6fl_seq_show() reader dereferences freed memory. There is no public exploit identified at time of analysis, and EPSS is 0.02% (7th percentile), consistent with a low-value local denial-of-service bug rather than a widely exploited flaw.
Local privilege escalation via a use-after-free race in the Linux kernel's Open vSwitch (openvswitch) module allows a local user with network-configuration capability to corrupt kernel memory. The flaw lives in ovs_netdev_tunnel_destroy(), which drops the tunnel netdev reference during teardown; when this runs after NETDEV_UNREGISTER has already detached the device, it races concurrent RCU readers that still observe vport->dev, producing a premature reference drop and stale-pointer access. The fix defers the netdev_put to the RCU release callback (vport_netdev_free), matching the non-tunnel path. There is no public exploit identified at time of analysis, and EPSS is negligible (0.02%, 7th percentile).
Out-of-bounds kernel memory access in the Linux kernel's network emulation scheduler (sch_netem) allows a local, privileged user to corrupt kernel memory and likely panic the system. When a netem qdisc has packet corruption enabled and an AF_PACKET TX_RING transmits fully non-linear packets over an IPIP tunnel, skb_headlen(skb) is 0, causing get_random_u32_below(0) to return an unconstrained 32-bit offset that indexes past the linear skb->data area. There is no public exploit identified at time of analysis, EPSS risk is very low (0.02%), and it is not on CISA KEV; it is a locally-triggerable memory-safety bug rather than a remotely exploitable one.
Kernel memory disclosure in the Linux kernel's AF_UNIX subsystem allows a local, low-privileged user to read stale or freed VFS data via the UNIX_DIAG_VFS netlink attribute. The sock_diag handler read a UNIX socket's u->path inode and device numbers without holding unix_state_lock, racing against unix_release_sock() which clears u->path under that lock before dropping the path reference. The fix reads the inode/device under unix_state_lock and emits the netlink attribute after unlocking; no public exploit identified at time of analysis and EPSS is negligible (0.02%).
Denial of service in the Linux kernel's rxrpc (AF_RXRPC) subsystem allows remote attackers to disrupt service-side RxRPC connections by sending duplicate or late RESPONSE packets that re-run the connection setup path. The flaw stems from an unlocked, post-transition state test that permitted RESPONSE processing outside the RXRPC_CONN_SERVICE_CHALLENGING window; the fix gates processing on that state under state_lock. No public exploit has been identified and EPSS probability is very low (0.02%), but any kernel exposing an rxrpc service (e.g., kafs/AFS) is potentially reachable over the network without authentication.
OS command injection in Linksys MR9600 router firmware 2.0.6.206937 allows authenticated administrators to execute arbitrary system commands via crafted 'pin' parameter to the BTRequestGetSmartConnectStatus JNAP action handler. Publicly available exploit code exists (CVSS E:P), enabling remote compromise of router with full system-level access. Vendor notified but unresponsive, leaving users without confirmed patch. EPSS data not available; CVSS 7.3 severity reflects high impact limited by high privilege requirement (PR:H).
Kernel memory disclosure and denial of service in the Linux kernel's openvswitch (OVS) datapath arises because validate_set() accepted OVS_KEY_ATTR_MPLS as a variable-length payload for SET/SET_MASKED actions while the action handler expects the fixed-size struct ovs_key_mpls. A local low-privileged actor able to install OVS flow actions can supply an undersized or oversized MPLS key to trigger an out-of-bounds read (rated C:H, I:N, A:H), leaking adjacent kernel memory or crashing the datapath. There is no public exploit identified at time of analysis, and EPSS is very low (0.02%, 7th percentile), consistent with a niche, privilege-gated networking bug rather than a mass-exploitation candidate.
Out-of-bounds kernel memory access in the Linux kernel's IPv6 netfilter ip6t_rt (Routing Header) match module allows a local user holding CAP_NET_ADMIN to install a crafted ip6tables rule whose addrnr field exceeds IP6T_RT_HOPS, causing rt_mt6() to read past the bounds of rtinfo->addrs[]. This can leak adjacent kernel memory (information disclosure) or crash the system (denial of service). No public exploit identified at time of analysis, and EPSS exploitation probability is negligible (0.02%, 7th percentile).
SQL injection in KLiK SocialMediaWebsite 1.0.1 and earlier allows remote unauthenticated attackers to execute arbitrary SQL commands via the c_id parameter in /includes/get_message_ajax.php. The vulnerability targets the private message handling component and permits unauthorized database access, modification, and potential data exfiltration. CVSS 7.3 reflects network-accessible, low-complexity exploitation requiring no authentication or user interaction, with partial impact to confidentiality, integrity, and availability. No public exploit code or CISA KEV listing identified at time of analysis, suggesting limited observed exploitation despite the accessible attack surface.
Improper verification of cryptographic signatures in Cesanta Mongoose versions up to 7.20 allows remote attackers to bypass GCM authentication tag validation in the mg_aes_gcm_decrypt function. The vulnerability has high attack complexity and requires no user interaction, but provides only integrity impact (not confidentiality or availability). Publicly available exploit code exists, and vendor has released patched version 7.21.
Command injection in Tenda F453 firmware up to version 1.0.0.3 allows authenticated remote attackers to execute arbitrary system commands via the TendaTelnet function in the /goform/telnet endpoint. The vulnerability has publicly available exploit code and may be actively used against deployed devices. Attack requires low-privilege authentication but carries significant risk due to the telnet service's direct command execution capability.
Server-side request forgery (SSRF) in devlikeapro WAHA up to version 2026.3.4 allows authenticated remote attackers to forge requests from the server via the media.controller.ts API endpoint, enabling potential reconnaissance, internal resource access, and lateral movement attacks. Publicly available exploit code exists and the vendor has not responded to disclosure efforts.
Server-side request forgery (SSRF) in AiraHub2 allows authenticated remote attackers to manipulate the connect_stream_endpoint and sync_agents functions in AiraHub.py, enabling arbitrary HTTP requests to internal or external systems. The vulnerability affects multiple endpoints and has publicly available exploit code; however, the vendor has not responded to disclosure attempts and uses a rolling release model, making patch status unclear.
Server-side template injection in AstrBot Dashboard API (version 4.22.1 and earlier) allows remote authenticated attackers with high privileges to execute arbitrary template code via the create_template function, leading to information disclosure and potential code execution. Publicly available exploit code exists, and the vendor has not yet responded to disclosure despite early notification.
Server-side request forgery (SSRF) in Pagekit up to version 1.0.18 allows authenticated high-privilege administrators to manipulate the url parameter in the /index.php/admin/system/update/download endpoint, enabling them to force the server to make arbitrary HTTP requests to internal or external systems. Publicly available exploit code exists, and the vendor did not respond to early disclosure efforts.
SQL injection in JiZhiCMS up to version 2.5.6 allows authenticated high-privileged administrators to execute arbitrary SQL queries via the sqls parameter in the /index.php/admins/Sys/addcache.html endpoint. The vulnerability is remotely exploitable and publicly available exploit code exists, though the low CVSS score (4.7) reflects the requirement for high-level administrative authentication, limiting real-world attack surface.
Stored cross-site scripting (XSS) in projeto-siga SIGA 11.0.3.18 allows authenticated remote attackers to inject malicious scripts via the Nome/Descrição parameter in the /sigawf/app/responsavel/novo endpoint. Successful exploitation requires user interaction (UI:R) and an authenticated session (PR:L), limiting impact to information disclosure (I:L). Public exploit code is available, though exploitation remains constrained by authentication and user interaction requirements.
Remote command injection in PicoClaw Web Launcher Management Plane (versions up to 0.2.4) allows unauthenticated attackers to execute arbitrary system commands via the /api/gateway/restart endpoint. CVSS 7.3 (AV:N/AC:L/PR:N/UI:N) indicates network-accessible exploitation without authentication. Proof-of-concept code exists (CVSS:E:P). Vendor has not responded to responsible disclosure (reported via GitHub issue #2307), indicating no official patch is available. The Web Launcher Management Plane component suggests this affects administrative/control interfaces, making it a high-priority target for internet-exposed deployments.
Denial-of-service in the Linux kernel's traffic control `act_csum` module allows a local low-privileged user to crash the kernel by exploiting missing VLAN header validation in `tcf_csum_act()`. Systems running any Linux kernel version from the introducing commit through the stable patch releases (pre-6.12.83, pre-6.18.24, pre-6.19.14) and configured with the `act_csum` tc action on paths carrying nested VLAN-tagged traffic are vulnerable. No public exploit has been identified and EPSS stands at 0.02% (5th percentile), indicating minimal real-world exploitation probability, but the issue is patch-confirmed across multiple stable branches.
Out-of-bounds array access in the Linux kernel's netfilter xt_multiport module enables a local attacker with low privileges to crash the system via a crafted malformed iptables rule. The checkentry validation path for multiport v1 rules fails to verify range encoding, allowing ports_match_v1() to step past the end of the ports[] array when processing rules where the last slot is flagged as a range start or when two consecutive range-start markers appear without an intervening range-end. Exploitation is confined to local actors with CAP_NET_ADMIN capability; no public exploit has been identified and EPSS sits at 0.02% (5th percentile), consistent with the narrow attack surface.
Denial of service in the Linux kernel's AF_ALG crypto socket subsystem allows a local, low-privileged user to crash or hang the system via crafted recvmsg calls. The flaw in af_alg_get_rsgl() causes receive buffer accounting to diverge from actual data attached to skcipher requests - when insufficient RX space remains, the kernel rounds the request length to zero rather than rejecting it, producing an invalid crypto operation. No exploitation has been confirmed (not in CISA KEV) and EPSS is 0.02%, reflecting the narrow local-only attack surface.
Code injection in Envoy up to 1.33.0 via improper query parameter handling in the Header Mutation filter allows authenticated remote attackers to inject arbitrary code through the params.add function, resulting in limited confidentiality and integrity impact. The CVSS 5.3 score reflects the requirement for prior authentication and limited scope of impact, though the injection vector in a core HTTP filtering component warrants prompt patching.
Soft-deleted public books in note-mark allow unauthenticated access to notes and assets via direct API endpoints and slug URLs. When a note-mark owner deletes a public book, the GORM soft-delete mechanism fails to filter raw SQL JOIN clauses in note and asset queries, leaving notes and uploaded content readable to any caller who knows the note ID or slug path. CVSS 5.3 (network, low complexity, no authentication required) reflects confidentiality impact; patch is available from vendor.
SQL injection in ShowDoc API Page Sort Endpoint allows authenticated remote attackers to manipulate the pages parameter and execute arbitrary SQL queries with limited confidentiality, integrity, and availability impact. Affected versions include 2.10.10, 3.6.2, and 3.8.0; vendor has released patch v3.8.1 but explicitly stated no backports will be provided for older versions.
SQL injection vulnerability in Zod CUID Data Type Handler affects versions up to 4.3.6, allowing authenticated remote attackers to manipulate input validation logic in the regex component and execute arbitrary SQL queries. The vulnerability has been publicly disclosed with proof-of-concept code available; the vendor was contacted early but provided no response, and no patch has been issued as of analysis time.
Reflected cross-site scripting (XSS) in Datacom DM4100 version 1.3.6.1.4.1.3709 allows remote attackers with high administrative privileges to inject malicious scripts via the VLAN Name parameter on the VLAN Page, exploitable only with user interaction. Public exploit code is available, and the vendor has not responded to early disclosure notifications.
Cross-site scripting (XSS) in Datacom DM4100 version 1.3.6.1.4.1.3709 allows authenticated remote attackers to inject malicious scripts via manipulation of the Name parameter in the Ethernet Configuration Page. The vulnerability requires high-privilege administrative access and user interaction to trigger, limiting real-world impact despite public exploit availability. Vendor has not responded to disclosure.
Cross-site scripting (XSS) vulnerability in BIVOCOM TR321 version 21.1.1.50 allows authenticated remote attackers with high privileges to inject malicious scripts via manipulation of the Network Name SSID parameter in the Wireless Setting component. The vulnerability requires user interaction to trigger and has limited integrity impact. Exploit code has been publicly published, though the vendor has not responded to disclosure attempts.
Reflected cross-site scripting (XSS) in BDCOM P3310D 0.4.2 Build 86345 via the Owner parameter on the New RMON Statistics Page allows authenticated remote attackers to inject malicious scripts. The vulnerability requires user interaction (clicking a crafted link) and high-privilege access to exploit, limiting real-world risk despite public exploit availability. The vendor has not responded to disclosure attempts.
Reflected cross-site scripting (XSS) in BDCOM P3310D 0.4.2 build 10.1.0F-86345 allows authenticated remote attackers with high privileges to inject malicious scripts via the Owner parameter on the New RMON History Page, requiring user interaction to execute. Publicly available exploit code exists, and the vendor has not responded to early disclosure notification.
Stored cross-site scripting (XSS) in BDCOM P3310D 0.4.2 build 86345 allows authenticated remote attackers to inject malicious scripts via the Description parameter in the rmon event Tab component. The vulnerability requires high-privilege authentication and user interaction to trigger, limiting practical exploitation scope. Public exploit code is available; however, the low CVSS score (2.4) reflects the authentication barrier and limited confidentiality/availability impact.
Reflected cross-site scripting (XSS) in BDCOM P3310D 0.4.2 10.1.0F Build 86345 allows high-privileged authenticated users to inject malicious scripts via the User name parameter in the New User Page (/index.asp), affecting only the injecting user's session due to session-scoped impact. Public exploit code is available, but exploitation requires administrative credentials and user interaction (form submission), significantly limiting real-world attack surface despite the remote attack vector.