| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Cocos AI is a confidential computing system for AI. In version 0.8.2 and earlier, Cocos AI accepts trusted execution environment attestation Evidence when firmware patch levels and software measurements match, without binding the Evidence to an allowed platform owner or unique machine identity and without enforcing revocation. A remote attacker who controls or compromises another platform trusted by the same verifier can substitute valid Evidence from that platform and cause authorization of the wrong platform. This vulnerability is addressed in 0.9.0. However, this vulnerability can't be patched in source code alone, since the vulnerability arises from the confidential computing threat model itself. |
| HomeBox is a home inventory and organization system. Prior to 0.27.0, the maybeUrl() utility accepts markdown-style links in entity custom-field values without restricting the URL scheme, so an authenticated user can store a javascript: URI that is rendered as a clickable anchor on an item page. When another authenticated user clicks the link, attacker-controlled JavaScript executes in the victim's HomeBox origin, enabling session theft, account takeover, or data exfiltration. This issue is fixed in version 0.27.0. |
| ntopng is a web-based network traffic monitoring application. Prior to 6.7.260717, scripts/lua/rest/v2/add/host/to_scan.lua and scripts/lua/rest/v2/exec/host/schedule_vulnerability_scan.lua accept scan_ports from any authenticated session, including non-administrator users, while deployments with login disabled expose the same request without authentication. The endpoints validate scan_ports with validateSingleWord, which permits shell metacharacters, and persist the value for the minute vulnerability-scan scheduler. The scheduler concatenates the value into an nmap command through nmap_scan_host and executes it through ntop.execCmd or ntop.execCmdAsync using popen. For the tcp_portscan and cve modules, a preliminary nmap -sn probe must first report the target as up before the attacker-controlled ports are used. When nmap is installed, this command injection allows arbitrary operating-system commands to execute as the ntopng service account without further user interaction. This issue is fixed in version 6.7.260717. |
| In Secret Server on-prem versions earlier than 12.3.1, Secret Server did not always validate the PIN or biometric check when using a FIDO2-compliant device |
| In Secret Server on-prem versions earlier than 12.3.1, a user under certain conditions could submit an unsigned JWT and receive cloudadmin permissions in response. |
| In Secret Server on-prem versions earlier than 12.3.1, a user could cause a two-factor pin code to be generated for any account without prior authentication, then bypass the initial login step and authenticate using the second factor alone. |
| Joomla Extension - balbooa.com - Unauthenticated arbitrary file deletion via path traversal in Gridbox < 2.20.5.0 |
| Joomla Extension - balbooa.com - Unauthenticated deletion of submission/field files in Gridbox < 2.20.5.0 |
| Joomla Extension - balbooa.com - Stored XSS in reviews via inconsistent sanitization in Gridbox < 2.20.5.0 |
| Joomla Extension - balbooa.com - Unauthenticated stored XSS in the "Recent Comments / Recent Reviews" widget in Gridbox < 2.20.5.0 |
| Joomla Extension - tassos.gr - Privileged stored XSS in Advanced Custom Fields 2.8.2-2.8.15, 3.0.0-3.1.8 and Smile Pack 1.1.2-1.2.9, 2.0.0-2.1.6 - A user who could edit a Map marker label, like an article author, could add a script that ran for anyone who opened the marker. |
| The Symfony HtmlSanitizer component does not apply the media URL policy to the srcset and imagesrcset attributes, allowing values that point to hosts or schemes not permitted by allowMediaSchemes() or allowMediaHosts() to bypass it, so that browsers load images from disallowed hosts in Symfony versions 6.1.x prior to 6.4.50, 7.x prior to 7.4.23, and 8.x prior to 8.1.11.The srcset attribute on img and source elements and the imagesrcset attribute on link elements are allowed by the sanitizer, but their URL candidates are never validated, so their values are kept verbatim while the same URL in src is dropped. The fixes for CVE-2026-45753 and CVE-2026-48761 extended the same attribute list but did not cover these two attributes. For example, with allowMediaSchemes(['https']) and allowMediaHosts(['trusted.example']), <img srcset="https://evil.example/a.png 1x"> survives sanitization, and the browser fetches the selected candidate as it would fetch src. Candidates with a scheme forbidden by the configuration, such as http, are kept as well.
Any application that sanitizes untrusted HTML and relies on the media schemes and hosts allowlists to restrict where images can be loaded from is affected. A single crafted value is enough to make the browser of a user viewing the sanitized content send a request to an attacker-controlled host, which can be used to track viewers or log their IP addresses. No authentication is needed if the application accepts HTML from unauthenticated users. |
| Incus is a system container and virtual machine manager. Prior to version 7.2.0, missing authorization checks exist for instance copying where an attacker knowing the name of a project that they don't have access to and the name of an instance in that project can copy the instance to a new project. This issue could allow an attacker to access secrets in instances they are not authorized to access. Version 7.2.0 patches the issue. |
| Processing a malformed RRC message leads to an out-of-bounds read, resulting in a denial-of-service condition. |
| A length validation flaw in gdbserver's agent-expression parser allows an unauthenticated remote attacker to read uninitialized heap memory from the gdbserver process and cause a denial of service. Disclosed heap content includes addresses suitable for defeating ASLR in further exploitation. |
| OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, OpenHuman places names, descriptions, titles, and parameter definitions returned by configured external MCP servers into the agent model context without neutralizing instructions or bounding the descriptive fields. The McpRemoteTool records produced by list_tools preserve this remote content, while tool filtering considers the tool name and the prompt-injection detector screens user messages rather than MCP definitions. A malicious or compromised MCP server can therefore return a poisoned description that is evaluated on every agent turn and may direct the model to invoke other available tools, disclose accessible data, or perform unauthorized actions without a separate malicious user message. The issue affects users who configure an attacker-controlled or compromised MCP server. This issue is fixed in version 0.57.13. |
| OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.56.0, the local server exposes the server-sent events endpoint at /events without bearer authentication, permits cross-origin browser access, and uses only a caller-supplied client_id to select WebChannelEvent objects from the process-wide event bus. An unauthenticated local client, or an attacker-hosted web page opened by the victim, can subscribe to another client's stream by supplying a known or guessed identifier; the fixed system identifier makes the proactive-message stream predictable. The exposed events can include complete assistant responses, streamed message content, tool names, tool arguments, and raw tool outputs intended for the local user interface. This issue is fixed in version 0.56.0. |
| OpenHuman is an open-source agent harness available as a desktop application and Rust library. In affected versions before 0.57.5, the approval gate uses the JSON-RPC bearer token as an approval session identifier. This exposes the credential through persisted approval records, events, debug logs, and RPC responses, including during normal desktop operation. An attacker who obtains these artifacts can recover the token and access the core RPC interface while it remains valid. This can expose sensitive agent and integration data and permit privileged operations. Stable tokens remain usable until rotated; launch-specific tokens remain usable while the corresponding core instance is active. The fix is included in version 0.57.5. |
| OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, a remote direct message received through a configured Yuanbao channel is converted into a generic channel AgentTurnRequest that uses the default tool policy without establishing an interactive approval context. Under the default Supervised autonomy policy with Node enabled, node_exec is classified as a write operation that should prompt for approval, but ApprovalGate treats missing APPROVAL_CHAT_CONTEXT as a non-interactive turn and allows execution. Because Yuanbao direct messages are open by default, an external sender who can message the connected bot can influence the agent to execute JavaScript and operating-system commands as the OpenHuman user without an approval prompt. The issue requires a configured Yuanbao channel, the default approval-requiring policy state, Node availability, and successful model selection of node_exec. This issue is fixed in version 0.57.13. |
| OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.56.0, the core server protects HTTP POST requests to /rpc with bearer middleware but exposes Socket.IO rpc:request and chat:start handlers without equivalent transport or handler-level authentication. A local unauthenticated client can connect to the Socket.IO namespace and dispatch requests through the same protected controller path that rejects a missing bearer on /rpc. Depending on the selected method, the attacker can read configuration or health data, start agent flows, interact with integrated account APIs, alter local data, or stop and restart services. This issue is fixed in version 0.56.0. |
