EUVD-2026-42621

Open WebUI: ReDoS in skill-mention regexes causes whole-instance DoS on default config...

EUVD-2026-42357

LiteLLM: Arbitrary file write via path traversal in Skills archive extraction...

CVE-2026-62225

OpenClaw versions before 2026.5.18 contain an authorization bypass vulnerability in skill command dispatch that allows lower-trust callers to execute or persist actions beyond their intended authorization. Attackers can bypass tool policy restrictions through configured input paths to perform...

CVE-2026-62225 OpenClaw < 2026.5.18 Authorization Bypass via Skill Command Dispatch

OpenClaw versions before 2026.5.18 contain an authorization bypass vulnerability in skill command dispatch that allows lower-trust callers to execute or persist actions beyond their intended authorization. Attackers can bypass tool policy restrictions through configured input paths to perform...

EUVD-2026-45113

OpenClaw versions before 2026.5.18 contain an authorization bypass vulnerability in skill command dispatch that allows lower-trust callers to execute or persist actions beyond their intended authorization. Attackers can bypass tool policy restrictions through configured input paths to perform...

CVE-2026-62225 OpenClaw < 2026.5.18 Authorization Bypass via Skill Command Dispatch

OpenClaw versions before 2026.5.18 contain an authorization bypass vulnerability in skill command dispatch that allows lower-trust callers to execute or persist actions beyond their intended authorization. Attackers can bypass tool policy restrictions through configured input paths to perform...

CVE-2026-62225

OpenClaw is affected by an authorization bypass in the skill command dispatch feature for versions before 2026.5.18. The vulnerability allows lower-trust callers to execute or persist actions beyond their intended authorization when the affected feature is enabled and reachable. The issue is docu...

PT-2026-60635

Name of the Vulnerable Software and Affected Versions OpenClaw versions prior to 2026.5.18 Description An authorization bypass exists in the skill command dispatch mechanism. This issue allows lower-trust callers to execute or persist actions that exceed their intended authorization levels. When...

Malicious code in xyq-drama-skill (PyPI)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 062ea591e5b995503deb15b174a37fc7af37cb5987a6c3d715d256fe0c3bfe10 setup.py registers a custom install cmdclass that, on pip install, downloads an opaque binary from...

CVE-2026-61443 PraisonAI before 1.6.78 Remote Code Execution via SkillTools

PraisonAI before 1.6.78 contains a remote code execution vulnerability in SkillTools.runskillscript that executes scripts without path containment validation. Attackers can supply absolute file paths to execute arbitrary scripts from any filesystem location, including those outside the intended...

CVE-2026-15629

A weakness has been identified in louisho5 picobot up to 0.2.0. Impacted is the function CreateSkill/GetSkill of the file internal/agent/tools/filesystem.go of the component Workspace Handler. Executing a manipulation can lead to link following. It is possible to launch the attack remotely. The...

EUVD-2026-43618

A weakness has been identified in louisho5 picobot up to 0.2.0. Impacted is the function CreateSkill/GetSkill of the file internal/agent/tools/filesystem.go of the component Workspace Handler. Executing a manipulation can lead to link following. It is possible to launch the attack remotely. The...

CVE-2026-15629 louisho5 picobot Workspace filesystem.go GetSkill link following

A weakness has been identified in louisho5 picobot up to 0.2.0. Impacted is the function CreateSkill/GetSkill of the file internal/agent/tools/filesystem.go of the component Workspace Handler. Executing a manipulation can lead to link following. It is possible to launch the attack remotely. The...

Malicious code in @spzhongwin/skill-logger-plugin (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 970980f38a6389d5b622ca41ba3b418f1538b2a38f595536c97acb934012b8b9 The package advertises itself as a skill usage/error logger but, when activated via the openclaw gatewaystart hook, opens a persistent WebSocket to a...

CVE-2026-15331 zhayujie CowAgent Skill Installation service.py _add_package path traversal

A vulnerability was identified in zhayujie CowAgent up to 2.1.0. The affected element is the function addurl/addpackage of the file agent/skills/service.py of the component Skill Installation Handler. The manipulation of the argument Name leads to path traversal. The attack may be initiated...

CVE-2026-15331

The CVE-2026-15331 vulnerability affects zhayujie CowAgent up to 2.1.0. The issue resides in the Skill Installation Handler’s component, specifically the _add_url/_add_package functions in agent/skills/service.py, where manipulation of the Name argument enables path traversal. A remote attack is ...

CVE-2026-59220

Open WebUI is an extensible, feature-rich, and user-friendly self-hosted AI platform. From 0.9.2 before 0.10.0, the SKILLMENTIONRE and stripre regular expressions in backend/openwebui/utils/middleware.py parsed skill mentions with overlapping quantifiers, allowing an authenticated chat message...

CVE-2026-59220 Open WebUI: ReDoS in skill-mention regexes causes whole-instance DoS on default config

Open WebUI is an extensible, feature-rich, and user-friendly self-hosted AI platform. From 0.9.2 before 0.10.0, the SKILLMENTIONRE and stripre regular expressions in backend/openwebui/utils/middleware.py parsed skill mentions with overlapping quantifiers, allowing an authenticated chat message...

CVE-2026-59220 Open WebUI: ReDoS in skill-mention regexes causes whole-instance DoS on default config

Open WebUI is an extensible, feature-rich, and user-friendly self-hosted AI platform. From 0.9.2 before 0.10.0, the SKILLMENTIONRE and stripre regular expressions in backend/openwebui/utils/middleware.py parsed skill mentions with overlapping quantifiers, allowing an authenticated chat message...

CVE-2026-59220

Open WebUI (self-hosted AI platform) has a ReDoS in theSkill mention handling. In versions before 0.10.0, the SKILL_MENTION_RE and strip_re regexes in backend/open_webui/utils/middleware.py parsed mentions with overlapping quantifiers, allowing an authenticated user to send a message containing ...