Loading story…
Palo Alto Networks Unit 42 reported that machine-learning-based traffic analysis identified malicious exploit activity that traditional intrusion prevention signatures can miss, particularly for zero-day and rapidly evolving attacks. The research said signature-based IPS remains effective for known threats but can lag when attackers introduce new payloads, variants, or previously undisclosed vulnerabilities, creating gaps that lead to false negatives and delayed protection.
The cases cited included command injection and SQL injection attempts tied to Atlassian Confluence CVE-2022-26134, Tenda AC18 CVE-2022-31446, Moodle CVE-2022-0332, Django CVE-2022-34265, an unidentified IoT zero-day affecting certain MIPS-based smart devices, and SQL injection payloads generated by sqlmap. Unit 42 said models trained on real-world benign and exploit traffic flagged these attacks with high confidence even where default IPS signature coverage was limited or unavailable, positioning machine learning as a complementary detection layer rather than a replacement for signatures.
See real exploitation activity before you spend the cycle.
3 events from the most recent confirmed update back to the earliest known activity.
The content describes an HTTP request observed on April 29, 2022 that triggered an alert and was assessed as a possible previously unknown attack targeting certain MIPS-based smart devices. The associated command-and-control server went offline shortly afterward, limiting verification of the suspected zero-day.
The content states that successful exploitation of Atlassian Confluence CVE-2022-26134 was observed in Cerber ransomware attacks. It presents this as a real-world example of command injection activity that its machine learning model could distinguish from benign traffic.
A Palo Alto Networks researcher discovered an exploit in the wild targeting Tenda AC18 CVE-2022-31446, a remote code execution flaw that allows arbitrary command execution on the device. The article says its machine learning model detected the exploitation attempt with high confidence.
See real exploitation activity behind this advisory so you can triage it against everything else in the queue.
1 reference tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Multiple weekly threat roundups and research posts reported a mix of active exploitation, new malware tradecraft, and breach claims. Ukraine’s CERT reported **APT28** rapidly weaponized a Microsoft Office zero-day (**CVE-2026-21509**) within roughly a day of Microsoft’s disclosure, using spearphishing emails with malicious DOC lures to deliver **Covenant** backdoors against Ukrainian government targets and EU-related entities. Separately, researchers described **ShadowHS**, a stealthy **fileless Linux** post-exploitation framework that runs in-memory (e.g., via `memfd`-style execution), uses encrypted multi-stage loading (AES-256-CBC), fingerprints defensive tooling (including major EDR agents), and retains operator-driven capabilities such as credential theft, lateral movement, and covert tunneling for exfiltration. Other reporting highlighted incident and exposure claims and defensive takeaways. Check Point described a **supply-chain compromise** affecting *eScan* (MicroWorld Technologies) in which malicious updates were pushed through the legitimate updater, prompting an emergency shutdown of global update services; it also noted **Crunchbase** confirmed a breach affecting **2M+ records** claimed by **ShinyHunters**, and cited extortion/leak claims involving **Qilin** (Tulsa International Airport) and **WorldLeaks** (Nike). Google’s legal/technical disruption of the **IPIDEA** residential proxy network was also cited as reducing available proxy nodes by millions and cutting off C2 domains used to route attacker traffic. Additional coverage described a phishing chain using a fake DHL invoice to abuse a signed Java utility via **DLL sideloading** (malicious `jli.dll`) and **process hollowing** into `AddInProcess32.exe` to run **Phantom Stealer**; detection-engineering updates emphasized new rules for Windows defense-evasion (e.g., tampering with Credential Guard/HVCI, disabling AMSI and the vulnerable driver blocklist) and expanded Kubernetes and Linux post-exploitation detections.
Jun 29, 2026
Multiple reports describe **social-engineering and supply-chain style attacks** that trick developers or AI-agent users into executing attacker-controlled instructions. North Korean operators have been linked to the **“Contagious Interview”** campaign, in which fake recruiter personas lure software developers into running “technical interview” projects that deploy malware such as **BeaverTail** and **OtterCookie** for credential theft and remote access; GitLab reported banning **131 related accounts** in 2025, with many repos using **hidden loaders** that fetched payloads from third-party services (e.g., *Vercel*) rather than hosting malware directly. Separately, OpenGuardrails reported a campaign on *ClawHub* (an OpenClaw AI agent “skills” repository) where attackers posted **malicious troubleshooting comments** containing Base64-encoded commands that download a loader from `91[.]92[.]242[.]30`, remove macOS quarantine attributes, and install **Atomic macOS (AMOS) infostealer**—a delivery method that can evade package-focused scanning because the payload is in comments, not the skill artifact. Research and incident writeups also highlight how **indirect prompt injection** and **malicious open-source packages** can compromise developer environments. NSFOCUS summarized a GitHub **MCP cross-repository data leak** scenario where attacker-injected instructions in public Issues could cause locally running AI agents to exfiltrate private repo data when agents act with broad GitHub permissions, and cited a similar hidden-command issue affecting an AI browser’s page summarization workflow. JFrog reported malicious npm packages (e.g., `eslint-verify-plugin`, `duer-js`) delivering multi-stage payloads including a **macOS RAT** (Mythic/Apfell) and a Windows infostealer, reinforcing ongoing risk from poisoned dependencies. In contrast, a DFIR case study on **CVE-2023-46604** exploitation of Apache ActiveMQ leading to **LockBit**-style ransomware, and a Medium post on recon/content-discovery techniques, are separate topics and not part of the AI-agent/developer social-engineering thread.
Jun 29, 2026
Security researchers and AI vendors reported rapid progress in applying frontier models to offensive and defensive cybersecurity work, including the discovery of **`CVE-2025-37899`**, a remote zero-day in the Linux kernel’s SMB implementation. A published account described using an advanced model to help identify the flaw, while OpenAI's **Daybreak** and Anthropic's **Project Glasswing** highlighted broader efforts to build AI systems for cyber operations, vulnerability research, and security workflows. SentinelLABS added evidence that these systems are moving beyond isolated tasks into sustained investigations, describing an eight-stage benchmark based on reverse engineering **fast16**, a 2005 Windows sabotage toolkit tied to interference with high-precision solvers used in nuclear-weapons modeling. In that evaluation, **GPT-5.6 Sol** was reported as the only publicly available model tested that completed the full long-horizon malware-analysis workflow, while other models showed useful point-in-time reasoning but failed to maintain consistency as new evidence forced earlier conclusions to be revised.
Jul 29, 2026