Don't Trust Us: A Privacy-By-Design Android Malware Detection Pipeline

Android malware detection increasingly relies on collecting and processing sensitive user data, including device identifiers, network artifacts, and runtime traces, while privacy is too often treated as a secondary concern. Existing privacy-aware approaches typically enforce privacy after data...

CVE-2026-44287

FastGPT is an AI Agent building platform. Prior to 4.15.0-beta1, the JavaScript sandbox worker at projects/code-sandbox/src/pool/worker.ts:356 blocks dynamic import with the regex /\bimport\s\/.testcode. JavaScript syntax accepts a block comment between import and ; the regex matches only ASCII...

SUSE-SU-2026:2204-1 Security update for busybox

This update for busybox fixes the following issue - CVE-2026-29004: a crafted DHCPv6 response can lead to a heap buffer overflow in the DHCPv6 client bsc1263989...

ovn: OVN: Information disclosure via crafted DHCPv6 packets

A flaw was found in OVN Open Virtual Network. A remote attacker, by sending crafted DHCPv6 Dynamic Host Configuration Protocol for IPv6 SOLICIT packets with an inflated Client ID length, could cause the ovn-controller to read beyond the bounds of a packet. This out-of-bounds read can lead to the...

ovn: OVN: Information disclosure via crafted DHCPv6 packets

A flaw was found in OVN Open Virtual Network. A remote attacker, by sending crafted DHCPv6 Dynamic Host Configuration Protocol for IPv6 SOLICIT packets with an inflated Client ID length, could cause the ovn-controller to read beyond the bounds of a packet. This out-of-bounds read can lead to the...

AgentRedBench: Dynamic Redteaming and Integration-Aware Defense for LLM Agents over SaaS Integrations

Indirect prompt injection in tool-use agents is a concrete production threat: LLM agents read from integrations third-party services such as Gmail, Salesforce, or Jira accessed through tool calls whose response content the user neither writes nor controls. Existing benchmarks under-measure the...

The vulnerability of the doCertificateUpdate() function in the LXD container management system allows a attacker to escalate their privileges.

The vulnerability of the doCertificateUpdate function in the LXD container management system is related to insufficient control over the modification of dynamically defined object properties. Exploiting this vulnerability could allow a malicious actor to increase their privileges remotely...

Defenses and Enablers for Skill Injection Attacks on Terminal Based Agents

Large language model LLM agents increasingly rely on reusable skills i.e. documents describing task-specific procedures. However, this introduces a new attack surface for agents to manage. We study two complementary directions for this threat. First, we evaluate guardian-based defenses: an...

CVE-2026-41897

Mantis Bug Tracker MantisBT is an open source issue tracker. From 1.0.0 to 2.28.1, lack of validation of filtertarget parameter on returndynamicfilters.php normally used as an AJAX in View Issues Page allows an attacker to inject arbitrary HTML if the target is a TEXTAREA custom field. This...

Nezha's authenticated DDNS webhook configuration allows blind SSRF from the dashboard host

Summary An authenticated Nezha dashboard user can create or update a DDNS profile with provider webhook and configure an arbitrary webhookurl, HTTP method, request body, and headers. When DDNS is triggered for a server that uses that profile, the dashboard process sends the configured request wit...

CVE-2026-44287

FastGPT is an AI Agent building platform. Prior to 4.15.0-beta1, the JavaScript sandbox worker at projects/code-sandbox/src/pool/worker.ts:356 blocks dynamic import with the regex /\bimport\s/.testcode. JavaScript syntax accepts a block comment between import and ; the regex matches only ASCII...

CVE-2026-9404

A vulnerability was identified in Totolink A8000RU 7.1cu.643b20200521. This affects the function setDdnsCfg of the file /cgi-bin/cstecgi.cgi of the component Web Management Interface. Such manipulation of the argument provider leads to os command injection. The attack may be launched remotely. Th...

CVE-2026-44287 FastGPT: sandbox escape to RCE - code-sandbox regex /\bimport\s*\(/ is bypassable

FastGPT is an AI Agent building platform. Prior to 4.15.0-beta1, the JavaScript sandbox worker at projects/code-sandbox/src/pool/worker.ts:356 blocks dynamic import with the regex /\bimport\s/.testcode. JavaScript syntax accepts a block comment between import and ; the regex matches only ASCII...

CVE-2026-44287

CVE-2026-44287 : In FastGPT, before 4.15.0-beta1, the JavaScript sandbox worker at projects/code-sandbox/src/pool/worker.ts:356 blocks dynamic import() with the regex /\bimport\s*(/. The payload import/**/("child_process") parses as valid dynamic import, escaping detection because the regex only ...

EUVD-2026-33431

FastGPT is an AI Agent building platform. Prior to 4.15.0-beta1, the JavaScript sandbox worker at projects/code-sandbox/src/pool/worker.ts:356 blocks dynamic import with the regex /\bimport\s/.testcode. JavaScript syntax accepts a block comment between import and ; the regex matches only ASCII...

CVE-2026-44287

FastGPT is an AI Agent building platform. Prior to 4.15.0-beta1, the JavaScript sandbox worker at projects/code-sandbox/src/pool/worker.ts:356 blocks dynamic import with the regex /\bimport\s/.testcode. JavaScript syntax accepts a block comment between import and ; the regex matches only ASCII...

CVE-2026-44287 FastGPT: sandbox escape to RCE - code-sandbox regex /\bimport\s*\(/ is bypassable

FastGPT is an AI Agent building platform. Prior to 4.15.0-beta1, the JavaScript sandbox worker at projects/code-sandbox/src/pool/worker.ts:356 blocks dynamic import with the regex /\bimport\s/.testcode. JavaScript syntax accepts a block comment between import and ; the regex matches only ASCII...

Improper Control of Dynamically-Managed Code Resources

Overview org.webjars.npm:vm2 is a sandbox that can run untrusted code with whitelisted Node's built-in modules. Affected versions of this package are vulnerable to Improper Control of Dynamically-Managed Code Resources through the WebAssembly.promising and WebAssembly.Suspending JSPI APIs in...

Improper Control of Dynamically-Managed Code Resources

Overview vm2 is a sandbox that can run untrusted code with whitelisted Node's built-in modules. Affected versions of this package are vulnerable to Improper Control of Dynamically-Managed Code Resources through the NodeVM constructor in lib/nodevm.js. An attacker can obtain host code execution by...

Improperly Controlled Modification of Dynamically-Determined Object Attributes

Overview vm2 is a sandbox that can run untrusted code with whitelisted Node's built-in modules. Affected versions of this package are vulnerable to Improperly Controlled Modification of Dynamically-Determined Object Attributes via the BaseHandler.set trap in lib/bridge.js. An attacker can mutate...