Attackers have shifted from hiding from AI tools to running inside them. By poisoning the config files that govern AI coding assistants, a new worm class achieves silent persistence, evades AI-based scanners, and spreads across an organization's repositories through developers' own tools. Key takeaways AI coding assistant configuration files, such as settings.json hooks, .cursorrules Cursor MDC rules, and similar harness files, are now explicit targets in supply-chain attacks, not collateral damage. These files simultaneously sit at the intersection of three trust relationships: The developer trusts them as config, the integrated development environment (IDE) executes them automatically, and the large language model (LLM) treats them as authoritative instructions, making them a uniquely powerful persistence vector. Defenders should treat harness config files as code that requires mandatory reviews and hash pinning in the CI/CD pipeline; enforce --ignore-scripts on package installs; and flag AI scanner refusals as a suspicious signal rather than a clean result. While analyzing a recent Mini Shai-Hulud campaign and its other variants, Tenable made a discovery: The payload wasn’t just stealing tokens and exfiltrating credentials. It was rewriting the configuration files that tell AI coding assistants what to do every time a developer opens a project. The Mini Shai-Hulud worm, which targets npm, PyPI, and other third-party registries simultaneously, includes a dedicated module that scans the developer’s home directory for configuration files belonging to AI coding tools by name, such as Anthropic’s Claude Code, Google’s Gemini CLI, Microsoft’s GitHub Copilot, SpaceX’s Cursor, OpenAI’s ChatGPT Codex, Cline’s eponymous tool, the open-source tool Aider, and others. For each one it finds, it injects a hook command which helps the malware stay persistent. From that point forward, every time the developer starts an AI coding session, the malware runs silently and automatically, and with the same level of trust the developer placed in their own tools. This is a deliberate architectural choice: The attacker modeled the execution environment, identified the highest-privilege auto-run mechanism available on a developer’s machine, and built the attack specifically for it. Supply-chain attacks used to end at credential theft. Now the AI agent’s harness is both the target and the propagation vector. Why AI agent harness files are a uniquely attractive target For years, defenders have focused supply-chain security on the package itself: Inspect the preinstall script, scan the tarball, verify the checksum. That model assumes the malicious payload lives only inside the installed artifact. The threat described here breaks that assumption: The payload lands in a completely different place. AI coding assistants’ configuration files occupy a structural position unlike any other file in a developer’s working environment. They are simultaneously trusted as developer configuration, so they survive aggressive code reviews focused on business logic; are executed automatically by the IDE, so no user interaction is required after initial compromise; and are treated as authoritative instructions by the AI model itself, so the LLM will act on their content without prompting the user. No other file class has all three properties at once. A malicious package.json script requires execution. A malicious comment in source code requires a developer to act on it. A malicious hook in .claude/settings.json runs every time the developer opens a session. The attack surface is also nearly invisible to conventional scanning. These files are small, structured, and look identical to legitimate developer config. A poisoned settings.json with a malicious SessionStart hook has the same schema and structure as a clean one; the only difference is what the command field points to. Meanwhile, the files are rarely subject to the same review discipline as production code. They live…
This content discusses the strategic integration of application security scanner data into exposure management platforms to improve visibility, prioritization, and remediation of code vulnerabilities. It highlights challenges posed by isolated application security data and accelerated code deployment, especially with AI-assisted coding increasing vulnerability rates. The integration enables security teams to contextualize vulnerabilities within the broader organizational risk, prioritize fixes effectively, and provide CISOs with actionable risk metrics. It is not a specific vulnerability or exploit but a security management approach to reduce organizational exposure to code flaws.
Yesterday, a reader reported to us a malicious ZIP archive (SHA256: a0104921a2d37ab87482ac9a9f5c3713479c118846c3e999178e75b81620c094[1]). Once unzipped, it contains a VHDX file that discloses a malicious JavaScript after being mounted (which is automatic on modern Windows OSs):
Microsoft has patched an actively exploited Exchange Server vulnerability that allows threat actors to execute arbitrary JavaScript code in cross-site scripting (XSS) attacks targeting Outlook Web Access users. [...]
This vulnerability allows remote attackers to bypass the sandbox on affected installations of OpenAI Codex. User interaction is required to exploit this vulnerability in that the target must use Codex to process a repository containing malicious JavaScript. The ZDI has assigned a CVSS rating of 8.6.
I spotted an interesting piece of JavaScript code that was delivered via a phishing email in a RAR archive. The file was called “cbmjlzan.JS” (SHA256:a8ba9ba93b4509a86e3d7dd40fd0652c2743e32277760c5f7942b788b74c5285) and is only identified as malicious by 15 AV's on VirusTotal[1].
The improper validation of user-supplied JavaScript code allows attackers to execute arbitrary code and access the file system. The post Critical Flowise Vulnerability in Attacker Crosshairs appeared first on SecurityWeek .
CrushFTP is a Java-based open source file transfer system. It is offered for multiple operating systems. If you run a CrushFTP instance, you may remember that the software has had some serious vulnerabilities: CVE-2024-4040 (the template-injection flaw that let unauthenticated attackers escape the VFS sandbox and achieve RCE), CVE-2025-31161 (the auth-bypass that handed over the crushadmin account on a silver platter), and the July 2025 zero-day CVE-2025-54309 that was actively exploited in the wild.
A critical sandbox escape vulnerability (CVE-2026-25049) in the n8n AI workflow automation platform allows attackers to bypass the sandbox's JavaScript sanitization and execute arbitrary commands on the server. This flaw arises from improper evaluation of JavaScript expressions within the sandbox, enabling potential full server compromise. Although no known exploits are currently reported in the wild, the high severity and ease of exploitation make this a significant threat. European organizations using n8n for automation workflows face risks of data breaches, service disruption, and unauthorized access. Immediate patching and sandbox hardening are essential to mitigate this vulnerability. Countries with high adoption of automation tools and strong digital infrastructure, such as Germany, France, and the UK, are particularly at risk. Defenders should prioritize restricting n8n server access, monitoring unusual activity, and applying vendor updates once available. Given the critical impact on confidentiality, integrity, and availability without requiring user interaction, this vulnerability is assessed as critical severity.
The protections against NPM supply chain attacks could be bypassed, leading to arbitrary code execution. The post ‘PackageGate’ Flaws Open JavaScript Ecosystem to Supply Chain Attacks appeared first on SecurityWeek .
The XSS vulnerability could allow remote attackers to execute arbitrary JavaScript code with administrator privileges. The post Ivanti EPM Update Patches Critical Remote Code Execution Flaw appeared first on SecurityWeek .
Chrome 143 stable was released with patches for 13 vulnerabilities, including a high-severity flaw in the V8 JavaScript engine. The post Chrome 143 Patches High-Severity Vulnerabilities appeared first on SecurityWeek .
The two bugs are high-severity type confusion and inappropriate implementation issues in the browser’s V8 JavaScript engine. The post Google Pays $100,000 in Rewards for Two Chrome Vulnerabilities appeared first on SecurityWeek .
SAP has rolled out additional protections for insecure deserialization bugs resolved in NetWeaver AS Java recently. The post SAP Patches Critical Vulnerabilities in NetWeaver, Print Service, SRM appeared first on SecurityWeek .
A routine monitoring by researchers uncovered an exploitation attempt on a honeypot server running TeamCity, a CI/CD tool. The attack exploited an exposed Java Debug Wire Protocol (JDWP) interface, leading to remote code execution, deployment of cryptomining payload, and establishment of multiple persistence mechanisms. The attack was notable for its rapid exploitation, use of a customized XMRig payload, and stealthy crypto-mining techniques. JDWP, designed for debugging Java applications, becomes a high-risk entry point when exposed to the Internet without proper authentication. The attackers used a structured sequence to achieve remote code execution, likely using a variant of jdwp-shellifier. They deployed a dropper script that installed an XMRig miner and set up various persistence mechanisms including boot scripts, systemd services, cron jobs, and shell configuration files.
