Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Your AI Code Assistant Can't Audit Itself. Here's Who Should.

AI coding tools make code generation fast and cheap, but self-attesting LLMs introduce critical security blind spots, secret leakages, and unverified vulnerabilities. In this executive panel, Mend.io leadership explores why the author cannot be the auditor and how independent, continuous verification protects modern applications without decreasing developer velocity. Visit us at Mend.io Key Topics Covered: Why AI-generated code suffers from higher vulnerability rates and volume spikes.

AI agents are making decisions before security can

AI agents can interpret objectives, select dependencies, write code, and even deploy before traditional security gates enter the process. The scanners may still run, but security could already be evaluating decisions it never had a chance to influence. That shift changes how AppSec teams need to think about agentic development.

Mend SCA now supports Bun, with reachability built in

TL;DR: Mend SCA now scans Bun-managed JavaScript and TypeScript projects, including both the text-based bun.lock and legacy binary bun.lockb lockfiles. It detects direct and transitive dependencies automatically, works in the Mend CLI and every repository integration, and supports reachability analysis so teams can fix the vulnerabilities their code actually calls first.

PHP2Shell is in CISA KEV: why runtime visibility matters

Five days after PHP shipped the fix, CISA added both PHP2Shell flaws to its Known Exploited Vulnerabilities (KEV) catalog. Federal civilian agencies have until October 29 to remediate. Last week’s question was whether attackers could turn the flaws into working shells. This week’s question is which of your images and hosts still run an affected PHP build, and who can fix each one.

Compromised LiteLLM 1.107.0 on PyPI Ships a Credential Stealer and Code Loader

On October 21, 2026, a malicious release of litellm, the Python library and proxy for calling LLM provider APIs, was published to PyPI as version 1.107.0. All seven compiled wheels carried three extra files that install straight into the root of site-packages: a.pth hook that runs a credential stealer every time Python starts, and a sitecustomize.py that starts a background loader built to fetch and run remote Python code every hour. The package has had more than 92 million downloads over the past month.

PHP2Shell: the fix is a runtime upgrade, not a code change

On October 21, 2026, PHP released 8.5.12, 8.4.27, 8.3.36, and 8.2.35 to close two critical remote code execution flaws in the PHP runtime. Both are rated CVSS v4.0 9.2. Together they go by the name “PHP2Shell.” The flaws sit in PHP itself, not in any application built on it. A secure codebase running on an affected PHP build is still exposed. That makes this an inventory problem first and a patching problem second.

ISO 42001 explained: Requirements, annexes & best practices

ISO/IEC 42001 is the first international standard for artificial intelligence management systems (AIMS). It provides a certifiable framework for organizations to develop, deploy, and govern AI systems responsibly. It helps you manage AI-specific risks, such as bias and security, and align with emerging regulations. Who should implement it.

Your AI agent is sandboxed. Its data still needs guardrails.

AI coding agents need access to real development context to be useful. They read repositories, modify files, execute commands, call tools, and communicate with external models. Docker Sandboxes gives those agents an isolated environment where they can work without unrestricted access to the developer’s host. But consider a simple scenario. An agent running entirely inside its sandbox reads a credential from a project configuration file and includes it in an otherwise legitimate request to its LLM.