MMSkillRisk: Can Agents Stay Safe When Multimodal Skills Become Traps?

Agent skills are shareable packages of procedural instructions, tools, and examples. Multimodal skills additionally include visual references that agents retrieve and inspect during execution. Because these images guide actions, attackers can disguise malicious instructions as ordinary visual...

Citrix NetScaler Improper Restriction of Operations within the Bounds of a Memory Buffer Vulnerability

Citrix NetScaler ADC and NetScaler Gateway contain an improper restriction of operations within the bounds of a memory buffer vulnerability that could allow for remote code execution or denial of service...

CVE-2026-100633 SiYuan 3.8.0 through 3.8.3 Path Traversal via MCP File Operations

SiYuan is a self-hosted personal knowledge management system. In versions 3.8.0 through 3.8.3, the MCP file tool's sensitive-path guard util.IsForbiddenAbsPath, invoked from resolvePath is applied only to the allowed root of recursive operations and not to each resolved descendant path — an...

CVE-2026-100589

OpenClaw versions before 2026.7.1 contain a sandbox bypass vulnerability in the browser tool that allows sandboxed sessions to access paired node browser actions despite allowHostControl=false configuration. Attackers with control over sandboxed agent input can select a paired node and perform ho...

CVE-2026-100589 OpenClaw before 2026.7.1 Sandbox Bypass via Browser Node

OpenClaw versions before 2026.7.1 contain a sandbox bypass vulnerability in the browser tool that allows sandboxed sessions to access paired node browser actions despite allowHostControl=false configuration. Attackers with control over sandboxed agent input can select a paired node and perform ho...

EUVD-2026-87462

OpenClaw versions before 2026.7.1 contain a sandbox bypass vulnerability in the browser tool that allows sandboxed sessions to access paired node browser actions despite allowHostControl=false configuration. Attackers with control over sandboxed agent input can select a paired node and perform ho...

CVE-2026-100589 OpenClaw before 2026.7.1 Sandbox Bypass via Browser Node

OpenClaw versions before 2026.7.1 contain a sandbox bypass vulnerability in the browser tool that allows sandboxed sessions to access paired node browser actions despite allowHostControl=false configuration. Attackers with control over sandboxed agent input can select a paired node and perform ho...

CVE-2026-100589 OpenClaw before 2026.7.1 Sandbox Bypass via Browser Node

OpenClaw versions before 2026.7.1 contain a sandbox bypass vulnerability in the browser tool that allows sandboxed sessions to access paired node browser actions despite allowHostControl=false configuration. Attackers with control over sandboxed agent input can select a paired node and perform ho...

CVE-2026-100589

OpenClaw versions before 2026.7.1 are vulnerable to a sandbox bypass in the browser tool. Despite an allowHostControl=false configuration, sandboxed sessions can access paired node browser actions. Attackers who control sandboxed agent input can select a paired node to perform host browser operat...

CVE-2026-100540 OpenClaw Feishu before 2026.8.1 Authentication Bypass via Disabled Account

OpenClaw Feishu before 2026.8.1 fails to validate whether a configured default account is disabled before selecting it for model tool operations. Attackers can exploit multi-account setups where a disabled default account retains credentials to read or modify Feishu resources through a revoked...

CVE-2026-100540 OpenClaw Feishu before 2026.8.1 Authentication Bypass via Disabled Account

OpenClaw Feishu before 2026.8.1 fails to validate whether a configured default account is disabled before selecting it for model tool operations. Attackers can exploit multi-account setups where a disabled default account retains credentials to read or modify Feishu resources through a revoked...

CVE-2026-100540: Incorrect Authorization

OpenClaw Feishu before 2026.8.1 fails to validate whether a configured default account is disabled before selecting it for model tool operations. Attackers can exploit multi-account setups where a disabled default account retains credentials to read or modify Feishu resources through a revoked...

CVE-2026-100589: Incorrect Authorization

OpenClaw versions before 2026.7.1 contain a sandbox bypass vulnerability in the browser tool that allows sandboxed sessions to access paired node browser actions despite allowHostControl=false configuration. Attackers with control over sandboxed agent input can select a paired node and perform ho...

PT-2026-99225

Name of the Vulnerable Software and Affected Versions OpenClaw versions prior to 2026.7.1 Description A sandbox bypass exists in the browser tool that allows sandboxed sessions to access paired node browser actions, even when the allowHostControl variable is set to false. An attacker who controls...

PT-2026-99425

Name of the Vulnerable Software and Affected Versions Kibana affected versions not specified Description A Confused Deputy issue in Kibana Agent Builder allows a non-administrative user with permissions to edit a shared agent to trigger privileged operations. These operations are executed using t...

PT-2026-99177

Name of the Vulnerable Software and Affected Versions OpenClaw Feishu versions prior to 2026.8.1 Description The software fails to validate if a configured default account is disabled before selecting it for model tool operations. In multi-account configurations, this allows attackers to use a...

CVE-2026-53627 GLPI: Unexpected access to update operations through the API

GLPI is a free asset and IT management software package. From 11.0.0 until 11.0.8, a low-privileged authenticated user can use the new API v2 to perform update operations that the same user is normally forbidden to perform through the user interface. The API update flow does not consistently...

CVE-2026-53627

The asset and IT management software GLPI versions 11.0.0 through 11.0.8 is vulnerable to an authorization bypass. A low-privileged authenticated user can exploit the API (v2) to perform update operations that are restricted through the standard user interface. This occurs because the API update ...

CVE-2026-53627 GLPI: Unexpected access to update operations through the API

GLPI is a free asset and IT management software package. From 11.0.0 until 11.0.8, a low-privileged authenticated user can use the new API v2 to perform update operations that the same user is normally forbidden to perform through the user interface. The API update flow does not consistently...

EUVD-2026-87273

GLPI is a free asset and IT management software package. From 11.0.0 until 11.0.8, a low-privileged authenticated user can use the new API v2 to perform update operations that the same user is normally forbidden to perform through the user interface. The API update flow does not consistently...