Credential Leakage in LLM Agent Skills: A Large-Scale Empirical Study

Third-party skills extend LLM agents with powerful capabilities but often handle sensitive credentials in privileged environments, making leakage risks poorly understood. We present the first large-scale empirical study of this problem, analyzing 17,022 skills sampled from 170,226 on SkillsMP usi...

SUSE-SU-2026:1170-1 Security update for perl-Crypt-URandom

This update for perl-Crypt-URandom fixes the following issue: Update to 0.550.0 0.55: - CVE-2026-2474: heap buffer overflow in the XS function crypturandomgetrandom bsc1258266. Changelog: - Fix for sysread/read failures. Thanks to Miha Purg for GH20. - Fix for test suite failures on STDOUT...

CVE-2026-34156 NocoBase Affected by Sandbox Escape to RCE via console._stdout Prototype Chain Traversal in Workflow Script Node

NocoBase is an AI-powered no-code/low-code platform for building business applications and enterprise solutions. Prior to version 2.0.28, NocoBase's Workflow Script Node executes user-supplied JavaScript inside a Node.js vm sandbox with a custom require allowlist controlled by WORKFLOWSCRIPTMODUL...

CVE-2026-34156 NocoBase Affected by Sandbox Escape to RCE via console._stdout Prototype Chain Traversal in Workflow Script Node

NocoBase is an AI-powered no-code/low-code platform for building business applications and enterprise solutions. Prior to version 2.0.28, NocoBase's Workflow Script Node executes user-supplied JavaScript inside a Node.js vm sandbox with a custom require allowlist controlled by WORKFLOWSCRIPTMODUL...

CVE-2026-34156

NocoBase exposes a sandbox escape in the Workflow Script Node: an attacker can traverse the sandbox through the host console object (console._stdout/console._stderr) prototype chain to reach the Function constructor, access process, require child_process, and achieve Remote Code Execution as root...

GHSA-PX3P-VGH9-M57C NocoBase Affected by Sandbox Escape to RCE via console._stdout Prototype Chain Traversal in Workflow Script Node

Summary NocoBase's Workflow Script Node executes user-supplied JavaScript inside a Node.js vm sandbox with a custom require allowlist controlled by WORKFLOWSCRIPTMODULES env var. However, the console object passed into the sandbox context exposes host-realm WritableWorkerStdio stream objects via...

NocoBase Affected by Sandbox Escape to RCE via console._stdout Prototype Chain Traversal in Workflow Script Node

Summary NocoBase's Workflow Script Node executes user-supplied JavaScript inside a Node.js vm sandbox with a custom require allowlist controlled by WORKFLOWSCRIPTMODULES env var. However, the console object passed into the sandbox context exposes host-realm WritableWorkerStdio stream objects via...

CVE-2026-32128

FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox fastgpt-sandbox includes guardrails intended to prevent file writes static detection + seccomp. These guardrails are bypassable by remapping stdout fd 1 to an arbitrary writable file descriptor using fcntl...

AWS VDP: Encryption context keys and values logged at INFO level

Component: cmd/server/main.go:101-106 Affected Version: aws-encryption-provider @ 4341c70 all versions Found by: Source audit TLP: TLP:Amber --- Summary The server startup code logs all encryption context key-value pairs at INFO level. Encryption context is metadata associated with KMS operations...

CVE-2026-32128

FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox fastgpt-sandbox includes guardrails intended to prevent file writes static detection + seccomp. These guardrails are bypassable by remapping stdout fd 1 to an arbitrary writable file descriptor using fcntl...

EUVD-2026-11408

FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox fastgpt-sandbox includes guardrails intended to prevent file writes static detection + seccomp. These guardrails are bypassable by remapping stdout fd 1 to an arbitrary writable file descriptor using fcntl...

CVE-2026-32128

FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox fastgpt-sandbox includes guardrails intended to prevent file writes static detection + seccomp. These guardrails are bypassable by remapping stdout fd 1 to an arbitrary writable file descriptor using fcntl...

CVE-2026-32128 FastGPT Python Sandbox Bypass of File-Write Restriction

FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox fastgpt-sandbox includes guardrails intended to prevent file writes static detection + seccomp. These guardrails are bypassable by remapping stdout fd 1 to an arbitrary writable file descriptor using fcntl...

CVE-2026-32128 FastGPT Python Sandbox Bypass of File-Write Restriction

FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox fastgpt-sandbox includes guardrails intended to prevent file writes static detection + seccomp. These guardrails are bypassable by remapping stdout fd 1 to an arbitrary writable file descriptor using fcntl...

PT-2026-24852

FastGPT is an AI Agent building platform. In 4.14.7 and earlier, FastGPT's Python Sandbox fastgpt-sandbox includes guardrails intended to prevent file writes static detection + seccomp. These guardrails are bypassable by remapping stdout fd 1 to an arbitrary writable file descriptor using fcntl...

ImageMagick's Security Policy Bypass through config/policy-secure.xml via "fd handler" leads to stdin/stdout access

The shipped “secure” security policy includes a rule intended to prevent reading/writing from standard streams: xml However, ImageMagick also supports fd: pseudo-filenames e.g., fd:0, fd:1. This path form is not blocked by the secure policy templates, and therefore bypasses the protection goal of...

GHSA-XWC6-V6G8-PW2H ImageMagick's Security Policy Bypass through config/policy-secure.xml via "fd handler" leads to stdin/stdout access

The shipped “secure” security policy includes a rule intended to prevent reading/writing from standard streams: xml However, ImageMagick also supports fd: pseudo-filenames e.g., fd:0, fd:1. This path form is not blocked by the secure policy templates, and therefore bypasses the protection goal of...

CVE-2026-25966

ImageMagick is free and open-source software used for editing and manipulating digital images. The shipped "secure" security policy includes a rule intended to prevent reading/writing from standard streams. However, ImageMagick also supports fd: pseudo-filenames e.g., fd:0, fd:1. Prior to version...

CVE-2026-25966

CVE-2026-25966 concerns ImageMagick’s security policy bypass via fd: pseudo-filenames (e.g., fd:0, fd:1). Prior to 7.1.2-15 and 6.9.13-40, the policy did not block this path form, allowing potential local access to stdin/stdout. A patch was added to more secure policies by default in 7.1.2-15/6.9...

Incomplete List of Disallowed Inputs

Overview Magick.NET-Q16-arm64 is a Magick.NET allows you can use ImageMagick without having to install ImageMagick on your server or desktop. More information about specific builds see the official docs https://github.com/dlemstra/Magick.NET/tree/main/docs Affected versions of this package are...