Behind the console: An AiTM phishing kit harvesting AWS console credentials and beyond

Datadog Security Labs dissects a June 2026 adversary-in-the-middle phishing kit that cloned the AWS console to harvest credentials and MFA tokens — a direct threat to any organisation managing cloud workloads via browser.

Lazarus Group's Latest: Brandjacking Campaign on npm

Lazarus Group escalates npm supply-chain attacks beyond typosquatting to brandjacking—mimicking legitimate package names via suffixes and version tricks to deliver second-stage payloads.

Europe Evolves Into Ransomware's Favorite Region

Dark Reading reports ransomware groups are pivoting toward EU targets and their supply chains after a global lull, making European enterprises — especially Mittelstand firms — front-line targets.

STOCKSTAY Another Day: The Latest Addition to Turla’s Intelligence Gathering Apparatus

Google's deep-dive into Turla's evolving .NET backdoor STOCKSTAY gives SOC teams fresh detection signatures for a Russia-linked APT that persistently targets European government and diplomatic networks.

The Gentlemen RaaS Uses GentleKiller EDR Framework Targeting 400 Security Processes

The Gentlemen RaaS operation distributes a modular EDR-killer framework called GentleKiller that targets roughly 400 security processes, reflecting a maturing affiliate ecosystem focused on disabling defences before encryption.

DragonForce Hackers Abuse Microsoft Teams Relays to Hide Backdoor.Turn C2 Traffic

DragonForce ransomware operators tunnel C2 traffic through Microsoft Teams relay infrastructure using a custom Go RAT called Backdoor.Turn, making malicious communications nearly indistinguishable from legitimate collaboration traffic.

Russian APT 'Gamaredon' Upgrades Its Arsenal, Requiring New Defenses

Russia's FSB-linked Gamaredon group has upgraded its malware delivery and C2 infrastructure concealment, requiring defenders to refresh detection signatures and hunting hypotheses for this prolific Ukraine-focused operation.

Turla group adds more malware to Russia’s espionage efforts against Ukraine

Google researchers detail StockStay, a fresh implant from Russia's Turla group, broadening the Kremlin's cyber-espionage toolkit aimed at Ukrainian targets — a reminder that state-level adversaries continually rotate tradecraft.

Five Eyes agencies sound alarm about AI’s threat to cybersecurity

Five Eyes agencies issued a joint alert warning that AI-enabled offensive capabilities are maturing within months rather than years, urging immediate defensive posture adjustments across critical sectors.

Amadey, StealC malware operations disrupted in Operation Endgame action

Europol-led Operation Endgame has dismantled Amadey and StealC infrastructure in a cross-border action involving Microsoft — a significant blow to the info-stealer ecosystem that feeds ransomware operators with stolen credentials.

FBI: Russian hackers now target Signal backup recovery keys

Russian intelligence operatives have evolved their Signal-targeting campaigns to steal backup recovery keys, granting access to full message histories — a significant escalation beyond real-time interception.

Node-gyp Supply Chain Compromise: A Self-Propagating npm Worm That Hides in binding.gyp

A novel npm worm abuses node-gyp's binding.gyp compilation step to execute code without lifecycle scripts—evading the standard mitigations that registries and lockdown tools rely on, while also persisting through GitHub token theft.

Miasma Malware Targets npm Packages and GitHub Actions in Supply Chain Attack

The Miasma malware family has expanded from npm into Go packages and now abuses GitHub Actions workflows, demonstrating how polyglot supply-chain attacks can propagate across multiple ecosystems simultaneously.

Compromised GitHub action codfish/semantic-release-action steals CI/CD secrets

Attackers hijacked codfish/semantic-release-action by retagging v2–v5 to a credential-stealing payload, exposing CI/CD secrets of any project that pinned to mutable version tags.

'Cordyceps': Mushrooming Malicious Pull Requests Threaten Developer Workflows

Malicious pull requests dubbed 'Cordyceps' exploit CI/CD workflow weaknesses in major projects including Azure Sentinel, Google AI Dev Kit, and Cloudflare Workers — a supply-chain threat at the build-pipeline level.

A Forgotten Contributor Account Compromised the Entire Mastra npm Package Scope

A dormant contributor account was hijacked to republish every package in the @mastra npm scope with an injected crypto-stealer dependency, illustrating how forgotten accounts with residual publish rights become the weakest link in scope ownership.

npm v12 delivers one of the biggest security improvements in years

npm v12 defaults install scripts to opt-in, effectively closing the most exploited execution vector behind this year's wave of supply-chain worms — a structural security improvement enterprise teams should fast-track.

Inherited Circuits, Learned Semantics: How Fine-Tuning Creates Evasion Vulnerabilities Invisible to Standard Evaluation

Fine-tuned LLM security classifiers can inherit hidden evasion paths invisible to standard test sets, meaning enterprises using fine-tuned models for malware or phishing detection may harbour blind spots.

Computer-Use and TOCTOU: What You Click Is Not What You Get!

Research reproduces a time-of-check/time-of-use flaw in AI agents with computer-use capabilities, showing that UI elements can be swapped between approval and execution to hijack automated actions.

FreeBSoD: Leveraging Language Models to Find and Exploit Kernel Bugs (Part 1 of 2)

Praetorian used Claude to discover and exploit FreeBSD kernel bugs — demonstrating that LLM-assisted vulnerability research is crossing from theory into practical 0-day output.

Interesting Paper Exploring Prompt Injection

Schneier highlights research showing LLMs learn to recognise instruction-block text styles rather than just tags, meaning prompt-injection defences built on role delimiters alone are fundamentally brittle.

Chai: Agentic Discovery of Cryptographic Misuse Vulnerabilities

Chai uses agentic AI to discover cryptographic misuse vulnerabilities — a bug class that resists traditional fuzzing — offering potential for automated crypto-hygiene audits in enterprise codebases.

Disposable Tooling: Building LLM-Generated Mythic Agents from Prompt to Deployment

SpecterOps demonstrates how LLMs can generate single-use Mythic C2 agents on demand, lowering the barrier for signature-evasive offensive tooling that defenders must now plan for.

Embedding Forbidden Text in Spyware to Discourage AI Analysis

Malware authors are embedding content about nuclear and biological weapons into spyware code to trigger AI safety guardrails and block automated analysis — a creative anti-analysis technique that undermines AI-driven threat detection.

Mandiant reveals how Cisco SD-WAN zero-day attacks gained root access

Mandiant details how attackers exploited CVE-2026-20245 in Cisco Catalyst SD-WAN to gain root and create rogue accounts — enterprises running affected appliances need immediate patching.

[UPDATE] [hoch] Microsoft Exchange: Mehrere Schwachstellen

BSI updates its high-severity Microsoft Exchange advisory covering admin-takeover, RCE, and spoofing — given Exchange's history as an APT target, European organisations should verify patch status urgently.

Probabilistic Agents in Deterministic Audits: Evaluating Multi-Agent Systems for Automated Audits Based on the German IT-Grundschutz

Companion study evaluates multi-agent LLM systems performing automated IT-Grundschutz audits, quantifying where probabilistic AI outputs clash with deterministic compliance requirements under NIS2.

An Approach for a Supporting Multi-LLM System for Automated Certification Based on the German IT-Grundschutz

Multi-LLM system with hybrid RAG automates large portions of BSI IT-Grundschutz certification — directly relevant to German Mittelstand organisations facing NIS2 compliance deadlines with limited audit resources.

AI and Liability

A German court ruling holds Google liable for AI-generated search summaries, rejecting the defence that users should verify AI outputs — setting a precedent that may shape EU AI Act liability interpretation.

Trump directs federal agencies to protect US data from quantum threats

A US executive order mandating federal post-quantum cryptography migration signals that European enterprises — especially those in transatlantic supply chains — should accelerate their own PQC readiness assessments now.

LastPass notifies users of yet another data breach