BIT-MONGODB-2026-1849 Mongod can run out of stack memory when expressions create deeply nested documents

MongoDB Server may experience an out-of-memory failure while evaluating expressions that produce deeply nested documents. The issue arises in recursive functions because the server does not periodically check the depth of the expression...

Server-side Request Forgery (SSRF)

Overview @langchain/community is a Third-party integrations for LangChain.js Affected versions of this package are vulnerable to Server-side Request Forgery SSRF via the RecursiveUrlLoader class. An attacker can access internal network resources or sensitive cloud metadata by supplying a public U...

GHSA-MPHV-75CG-56WG LangChain Community: redirect chaining can lead to SSRF bypass via RecursiveUrlLoader

Summary A redirect-based Server-Side Request Forgery SSRF bypass exists in RecursiveUrlLoader in @langchain/community. The loader validates the initial URL but allows the underlying fetch to follow redirects automatically, which permits a transition from a safe public URL to an internal or metada...

LangChain Community: redirect chaining can lead to SSRF bypass via RecursiveUrlLoader

Summary A redirect-based Server-Side Request Forgery SSRF bypass exists in RecursiveUrlLoader in @langchain/community. The loader validates the initial URL but allows the underlying fetch to follow redirects automatically, which permits a transition from a safe public URL to an internal or metada...

CVE-2026-27795

LangChain is a framework for building LLM-powered applications. Prior to version 1.1.8, a redirect-based Server-Side Request Forgery SSRF bypass exists in RecursiveUrlLoader in @langchain/community. The loader validates the initial URL but allows the underlying fetch to follow redirects...

CVE-2026-27795 LangChain Community: redirect chaining can lead to SSRF bypass via RecursiveUrlLoader

LangChain is a framework for building LLM-powered applications. Prior to version 1.1.8, a redirect-based Server-Side Request Forgery SSRF bypass exists in RecursiveUrlLoader in @langchain/community. The loader validates the initial URL but allows the underlying fetch to follow redirects...

CVE-2026-27795 LangChain Community: redirect chaining can lead to SSRF bypass via RecursiveUrlLoader

LangChain is a framework for building LLM-powered applications. Prior to version 1.1.8, a redirect-based Server-Side Request Forgery SSRF bypass exists in RecursiveUrlLoader in @langchain/community. The loader validates the initial URL but allows the underlying fetch to follow redirects...

LangChain.js 代码问题漏洞

LangChain.js is an open-source implementation of a context-aware reasoning application developed by LangChain. Versions of LangChain.js prior to 1.1.8 contained code vulnerabilities. These vulnerabilities stemmed from the RecursiveUrlLoader component, which allowed unauthorized access to...

CVE-2026-27115

ADB Explorer is a fluent UI for ADB on Windows. Versions 0.9.26020 and below have an unvalidated command-line argument that allows any user to trigger recursive deletion of arbitrary directories on the Windows filesystem. ADB Explorer accepts an optional path argument to set a custom data...

PortSwigger-DirectroyTraversal

PortSwigger Lab: File Path Traversal Non-Recursive Strip Bypa...

CVE-2026-27115

ADB Explorer is a fluent UI for ADB on Windows. Versions 0.9.26020 and below have an unvalidated command-line argument that allows any user to trigger recursive deletion of arbitrary directories on the Windows filesystem. ADB Explorer accepts an optional path argument to set a custom data...

ADB Explorer 安全漏洞

ADB Explorer is a ADB user interface developed by Alex B as an individual developer. Versions of ADB Explorer prior to 0.9.26020 contain security vulnerabilities, which stem from unvalidated command-line parameters and could lead to recursive deletion of directories...

PT-2026-21279

ADB Explorer is a fluent UI for ADB on Windows. Versions 0.9.26020 and below have an unvalidated command-line argument that allows any user to trigger recursive deletion of arbitrary directories on the Windows filesystem. ADB Explorer accepts an optional path argument to set a custom data...

Incomplete List of Disallowed Inputs

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Incomplete List of Disallowed Inputs in the tools.exec.safeBins configuration. An attacker can gain unauthorized access to the filesystem by leveraging allowed sort output flags -o or...

GHSA-4685-C5CP-VP95 OpenClaw safeBins stdin-only bypass via sort output and recursive grep flags

Summary tools.exec.safeBins could be bypassed for filesystem access when sort output flags -o / --output or recursive grep flags were allowed through safe-bin execution paths. Affected Packages / Versions - Package: openclaw npm - Affected versions: = 2026.2.19 - Latest published version at triag...

OpenClaw safeBins stdin-only bypass via sort output and recursive grep flags

Summary tools.exec.safeBins could be bypassed for filesystem access when sort output flags -o / --output or recursive grep flags were allowed through safe-bin execution paths. Affected Packages / Versions - Package: openclaw npm - Affected versions: = 2026.2.19 - Latest published version at triag...

Prototype Pollution

Overview devalue is a JSON.stringify, but handles cyclical references, repeated references, undefined, regular expressions, dates, Map and Set. Affected versions of this package are vulnerable to Prototype Pollution via the uneval method. An attacker can manipulate object prototypes by supplying...

PT-2026-26236

Summary tools.exec.safeBins could be bypassed for filesystem access when sort output flags -o / --output or recursive grep flags were allowed through safe-bin execution paths. Affected Packages / Versions - Package: openclaw npm - Affected versions: = 2026.2.19 - Latest published version at triag...

Recursive Language Models for Jailbreak Detection: A Procedural Defense for Tool-Augmented Agents

Jailbreak prompts are a practical and evolving threat to large language models LLMs, particularly in agentic systems that execute tools over untrusted content. Many attacks exploit long-context hiding, semantic camouflage, and lightweight obfuscations that can evade single-pass guardrails. We...

PT-2026-20517

MajorDoMo aka Major Domestic Module allows unauthenticated arbitrary module uninstallation through the market module. The market module's admin method reads gr'mode' from $ REQUEST and assigns it to $this-mode at the start of execution, making all mode-gated code paths reachable without...