Discovering the latest attacks and providing defensive measures to keep organizations safe

TeamPCP expands: Supply chain compromise spreads from Trivy to Checkmarx GitHub Actions

AI coding agents are running on your machines — Do you know what they're doing?

AI coding agents are now running on developer laptops and inside CI/CD pipelines across every sector. They write code, execute commands, read files, and make network connections, often without the developer watching. Unlike almost every other piece of software on those same machines, there is no established detection layer that understands what normal agent behavior looks like, let alone what an attack looks like at that level.

CVE-2026-33017: How attackers compromised Langflow AI pipelines in 20 hours

On March 17, 2026, a critical vulnerability was disclosed in Langflow, the open-source visual framework for building AI agents and Retrieval-Augmented Generation (RAG) pipelines. The vulnerability, CVE-2026-33017, is an unauthenticated remote code execution (RCE) in the public flow build endpoint that allows attackers to execute arbitrary Python code on any exposed Langflow instance, with no credentials required and only a single HTTP request to get moving.

Detecting CVE-2026-3288 & CVE-2026-24512: Ingress-nginx configuration injection vulnerabilities for Kubernetes

On March 9, 2026, the Kubernetes ingress-nginx project merged a fix for CVE-2026-3288 (CVSS 8.8 HIGH), a configuration injection vulnerability in the NGINX Ingress Controller. The vulnerability allows any user with permission to create or modify Ingress resources by inserting a double-quote character (“) into the Ingress path field. Because the field does not properly sanatize input, an attacker can break the expected syntax with a quotation mark and inject arbitrary nginx configuration directives into the generated configuration. The official advisory states that the vulnerability can lead to remote code execution (RCE) and the disclosure of secrets accessible to the controller.

LLMjacking: From Emerging Threat to Black Market Reality

Since its emergence in May 2024, LLMjacking has evolved from a novel security concern into an industrialized cybercrime marketplace. When this new class of cloud-focused AI attack was reported, researchers predicted that motivated actors would commercialize the practice. Now, additional investigations confirm those predictions: an underground marketplace is now actively monetizing unauthorized AI access at scale. LLMjacking is the new cryptomining.

AI-assisted cloud intrusion achieves admin access in 8 minutes

On November 28, 2025, the Sysdig Threat Research Team (TRT) observed an offensive cloud operation targeting an AWS environment in which the threat actor went from initial access to administrative privileges in less than 10 minutes. The attack stood out not only for its speed, but also for multiple indicators that suggest the threat actor leveraged large language models (LLMs) throughout the operation to automate reconnaissance, generate malicious code, and make real-time decisions.

VoidLink threat analysis: Sysdig discovers C2-compiled kernel rootkits

How threat actors are using self-hosted GitHub Actions runners as backdoors

Using Shai-Hulud as a case study, explore how attackers can abuse GitHub's self-hosted runner infrastructure to establish persistent remote access.

EtherRAT dissected: How a React2Shell implant delivers 5 payloads through blockchain C2

On December 8, the Sysdig Threat Research Team (TRT) reported that a possible North Korean-linked actor had deployed EtherRAT, a novel Ethereum-based implant, in React2Shell attacks. The malware goes beyond other React2Shell cryptomining attacks, blending command and control (C2) traffic into blockchain activity and aggressively harvesting credentials. Additionally, this blog marks the first time the React2Shell exploit has been publicly documented in active malware.

EtherRAT: DPRK uses novel Ethereum implant in React2Shell attacks

Sysdig Threat Research analyzes EtherRAT, a DPRK-linked Ethereum C2 implant exploiting React2Shell with multi-vector Linux persistence and evasion techniques.

Detecting React2Shell: The maximum-severity RCE vulnerability affecting React Server Components and Next.js

On December 3, 025, the React Team disclosed CVE-2025-55182, a critical unauthenticated remote code execution (RCE) vulnerability in React Server Components (RSCs). Dubbed "React2Shell", this maximum-severity flaw (CVSS 10.0) allows for potentially malicious code execution with a single crafted HTTP request. Next.js, which can be impacted downstream by React2Shell, has assigned CVE-2025-66478 to track its exposure.

Security briefing: November 2025

November is a month for reminiscing — looking back over the last 11 months to take stock of what’s been done in anticipation of what’s to come. This month, we’re reminded that “new” threats are not always novel; sometimes they emerge from the past, repurposed by threat actors who never stop innovating.

Return of the Shai-Hulud worm affects over 25,000 GitHub repositories

A newly evolved variant of the Shai-Hulud supply-chain worm is rapidly spreading through backdoored NPM packages, compromising nearly 1,000 packages and leaking credentials from more than 25,000 GitHub repositories since November 24, 2025. This version executes during NPM’s pre-install phase, using a lightweight dropper to install the Bun runtime and launch a 10-MB obfuscated payload that hunts for cloud secrets, GitHub tokens, and NPM credentials. When GitHub access is available, the worm creates a covert public repository with a randomly generated name and the description “Sha1-Hulud: The Second Coming,” then installs a hidden self-hosted GitHub Actions runner on the victim’s machine to maintain persistence and enable remote code execution. It also injects malicious workflows into accessible GitHub projects to extract repository secrets, exfiltrates all findings to the attacker-controlled repository, and removes evidence of its activity. If no NPM token is found, the malware becomes destructive, securely wiping all writable files in the user’s home directory.

Detecting CVE-2024-1086: The decade-old Linux kernel vulnerability that’s being actively exploited in ransomware campaigns

CVE-2024-1086, a decade-old Linux kernel vulnerability, is now being actively exploited in ransomware campaigns. This blog breaks down how attackers are weaponizing the flaw to gain root privileges, why so many systems remain exposed, and how Sysdig can help you detect exploitation before damage is done.

Hunting reverse shells: How the Sysdig Threat Research Team builds smarter detection rules

The Sysdig Threat Research Team (TRT) continuously analyzes attacker tactics and techniques, transforming those insights into effective detection rules for Sysdig customers and open source Falco users.

New runc vulnerabilities allow container escape: CVE-2025-31133, CVE-2025-52565, CVE-2025-52881

Three new runc vulnerabilities could let attackers escape containers and gain host access. Sysdig TRT explains the risks, affected versions, and how to detect and mitigate them with Sysdig Secure and Falco.

Security briefing: October 2025

October, also known in our community as Cybersecurity Awareness Month, was marked with a chilling reminder: pay attention to resilience and response speed. 

Understanding CVE-2025-49844: “RediShell” Critical Remote Code Execution in Redis

On October 3, 2025, CVE-2025-49844 was released, describing a critical remote code execution vulnerability in the widely used open-source in-memory data store, Redis. With a CVSS score of 10.0, this issue is very severe and should be addressed quickly.

Sysdig Security Briefing: September 2025

Last month, the NPM ecosystem was ablaze with hundreds of NPM packages compromised. After the first half of the month, every security researcher seemed to be hunting for issues in the NPM ecosystem, trying to find the next viral story. 

Shai-Hulud: The novel self-replicating worm infecting hundreds of NPM packages

A new supply chain attack against the NPM repository is using novel, self-propagating malware (also known as a worm) to continue spreading itself.

Detecting and Mitigating IngressNightmare – CVE-2025-1974

On Monday, March 24, 2025, a set of critical vulnerabilities affecting the admission controller component of the Ingress NGINX Controller…

tj-actions/changed-files with Falco Actions

A compromise (CVE-2025-30066) was discovered in the popular GitHub Action tj-actions/changed-files on March 14, 2025. It impacted tens of thousands…