Briefs

Two critical flaws in Coder's OIDC login mechanism, CVE-2026-55075, allow an attacker to achieve account takeover by exploiting email-based user matching without proper IdP subject checks and bypassing the `email_verified` claim, leading to full access to victim workspaces and resources.

A critical vulnerability, CVE-2026-55077, in the Coder platform allowed a user with the `user-admin` role to reset the password of an `owner` account without needing the current password via the `PUT /api/v2/users/{user}/password` endpoint, leading to privilege escalation and full deployment control.

A malicious or compromised Coder server can exploit CVE-2026-55427 to inject unsanitized SSH configuration values via `coder config-ssh` into developer workstations, enabling arbitrary code execution on client machines.

A high-severity vulnerability (CVE-2026-54641) in OpenRemote's `UserResourceImpl.java` allows a realm administrator in a multi-tenant deployment to perform cross-realm user enumeration and privilege-level reconnaissance by reading sensitive user information (profile, client roles, and realm roles) from any other realm, including the master realm, due to missing authorization checks in specific REST API endpoints.

A high-severity vulnerability, CVE-2026-54771, in Langroid applications allows untrusted users to directly invoke internal tools via raw JSON payloads, even when these tools are configured not to be used by the LLM, enabling malicious actors to bypass security controls and execute sensitive operations like file read/write, database queries, or access to internal orchestration tools.

Langroid is vulnerable to a critical Remote Code Execution (RCE) in its `TableChatAgent` and `VectorStore` components when `full_eval=True` due to CVE-2026-54769; the `eval()` function fails to properly scrub `__builtins__` from `globals`, allowing attackers to inject `__import__('os').system()` calls via crafted prompt payloads, leading to unauthenticated RCE, unauthorized data access, or system compromise on the host running the Langroid agent.

A critical vulnerability (CVE-2026-53486) in the `@xhmikosr/decompress` and unmaintained `decompress` npm packages allows attackers to craft malicious archives that, upon extraction, can write or read files outside the target directory, expose arbitrary file contents, or create setuid/setgid files leading to arbitrary file system modification, information disclosure, and potential privilege escalation.

flyto-core is vulnerable to unauthenticated command execution via its HTTP MCP endpoint (`POST /mcp`), allowing remote attackers to execute arbitrary OS commands with server privileges by invoking `sandbox.execute_shell` through JSON-RPC requests, potentially leading to full system compromise.

This brief details a common Linux technique where attackers use scripting interpreters to download malicious payloads from external sources and immediately pipe them into another interpreter for execution, often for purposes like persistence or data exfiltration.

The Scriban templating engine, specifically its `TypedObjectAccessor`, allows template code to write to arbitrary CLR object properties, including those with `private set`, `internal set`, and `init` modifiers, effectively bypassing intended C# access restrictions. This mass assignment (CWE-915) and access-modifier bypass (CWE-284) vulnerability can lead to unauthorized modification of sensitive host object properties, such as changing `user.is_admin = true`, with changes persisting after template rendering, affecting Scriban versions up to and including 7.2.1, with the `init` bypass specifically impacting .NET 5+.

An authenticated workflow author can bypass `flyto-core`'s Server-Side Request Forgery (SSRF) protection by crafting URLs with IPv6 transition addresses that embed private IPv4s, allowing for data exfiltration from internal services like cloud instance metadata.

This brief details a detection rule for the `xdg-open` command on Linux systems, which attackers abuse to trick users into opening malicious documents or URLs, leading to user execution and potential system compromise.

This brief details the detection of shell command execution initiated by the Elastic Endpoint agent on Linux systems, indicating potential post-exploitation activity such as remote access or command and control via misuse of the endpoint's response capabilities.

When Cilium L7 functionality is enabled, a world-accessible Envoy admin socket is inadvertently created on cluster nodes. This misconfiguration (CVE-2026-49445) allows a local attacker to gain unauthorized access to Envoy's administrative endpoints, leading to sensitive information disclosure, such as the exposure of TLS secrets, and significant cluster disruption, including the interruption of traffic and the termination of Envoy processes.

Formie Hidden fields in versions prior to 3.1.27 are vulnerable to Server-Side Template Injection (SSTI), allowing an unauthenticated attacker to inject Twig syntax into request-derived default values, potentially leading to remote code execution, sensitive information disclosure, or application state modification.

A privilege escalation threat in Google Kubernetes Engine (GKE) involves an attacker creating or modifying a pod to enable host Inter-Process Communication (IPC) namespace sharing, which exposes host IPC mechanisms and can lead to privilege escalation within the cluster by allowing the pod to interact directly with the underlying host's processes.

This brief details the Process Ghosting technique, an advanced evasion method leveraged by attackers to create and execute processes from files that no longer exist on disk, challenging traditional endpoint detection solutions and hindering forensic investigations.

A high-severity off-by-one error, CVE-2026-58380, in GIMP's PNM file format parser (specifically the `pnmscanner_gettoken()` function) allows an attacker to corrupt memory by crafting a malicious PNM file, potentially leading to denial of service or arbitrary code execution when the file is opened.

Adversaries utilize parent process ID (PPID) spoofing on Windows systems to create elevated child processes, typically to SYSTEM privileges, thereby evading process monitoring defenses and facilitating privilege escalation.

This brief details how attackers may leverage the `systemd-run` utility on Linux systems for defense evasion and execution by running commands as detached, transient services or scopes to obscure their activities and parent-child process chains.

This brief details the detection of Linux processes utilizing `curl` to contact known public IP address lookup web services, a common post-exploitation technique employed by malware and adversaries to ascertain a host's internet-facing IP, impacting reconnaissance and command-and-control tailoring.

Adversaries can leverage Living Off The Land Binaries (LoLBins) such as `openssl`, `python`, or `ruby` to load malicious shared object files (`.so`) into memory on Linux systems, aiming to evade detection by disguising the payload as legitimate process activity; detection engineers must investigate the full command line, parent process chain, executing user, and the reputation/location of the referenced shared object file to differentiate malicious activity from legitimate development or administration tasks.

Threat actors leverage C2 agents like Poseidon and Athena, operating from suspicious Linux writable directories, to establish network connections with C2 frameworks such as Mythic, subsequently creating files to stage further malicious activities.

This brief details the detection of DNS queries targeting commonly abused Remote Monitoring and Management (RMM) or remote access software domains, originating from non-browser processes, which is a common tactic for command and control, persistence, and lateral movement by threat actors.

Multiple vulnerabilities in OpenSSH versions prior to 10.4 allow attackers to bypass security policies, cause denial of service, and exploit other unspecified security issues, requiring users to update to OpenSSH 10.4 or later.

Multiple vulnerabilities, including Server-Side Request Forgery (SSRF), Cross-Site Scripting (XSS), and Denial of Service (DoS), have been discovered in Roundcube Webmail versions 1.6.x prior to 1.6.17 and 1.7.x prior to 1.7.2, allowing remote attackers to impact service availability and potentially execute malicious code or access internal resources.

A vulnerability, CVE-2026-54291, has been discovered in PostgreSQL JDBC versions 42.7.4 up to, but not including, 42.7.12, allowing an attacker to bypass security policies within applications utilizing the affected driver, potentially leading to unauthorized access or actions.

A critical PHP object injection vulnerability (CVE-2026-48909) in JoomShaper SP LMS versions <= 4.1.3 allows unauthenticated attackers to achieve remote code execution (RCE) via a crafted 'lmsOrders' cookie, leading to webshell deployment on vulnerable Joomla installations (< 5.2.2).

A public exploit has been published for a Stored XSS vulnerability in KeepInMind version 0.8.4.2, significantly increasing the risk for unpatched installations.

A local privilege escalation vulnerability (CVE-2026-36213) in MEmu Android Emulator 9.2.7.0 allows a low-privileged user to replace the 'MemuService.exe' binary due to insecure NTFS permissions, leading to arbitrary code execution with NT AUTHORITY\SYSTEM privileges upon service restart.