Intelligence Feed

2mo ago · hacker-news

Researcher Gareth Heyes from PortSwigger demonstrated novel CSS-based attacks that exploit rendering discrepancies in webmail clients to bypass security boundaries, enabling password theft, token leakage, and UI manipulation. The attacks affect multiple providers including Outlook, Gmail, Yahoo Mail, AOL, Fastmail, and Proton Mail, using techniques such as CSS mutation, label-jacking, paste-race conditions, and image-set() fallback abuse to exfiltrate sensitive data or spoof authentication interfaces. While some mitigations have been applied—such as Fastmail fixing CSS mutation bugs and Proton Mail proxy bypass no longer working—several vectors, including Outlook label-jacking and Gmail's image-set() bypass, remained exploitable at time of publication. The research highlights risks in AI-connected email systems, where injected CSS can manipulate AI tools like Anthropic's Cowork or OpenAI's Atlas to leak tokens or perform unintended actions.

2mo ago · hacker-news

A malicious Visual Studio Code extension named 'solidity-pro' has been identified as stealing cryptocurrency wallets, API keys, and credentials from developers. The threat delivers a browser wallet and credential stealer that exfiltrates sensitive data such as mnemonic phrases, SSH keys, GitHub tokens, AWS keys, and Telegram bot tokens via Telegram bots. The malware uses heavy obfuscation, delayed activation, and clean intermediate versions to evade detection by static scanners and sandbox environments.

2mo ago · hacker-news

Security researchers from PromptArmor and Varonis independently discovered vulnerabilities in Atlassian Rovo that allow attackers to exfiltrate Jira and Confluence data via maliciously crafted inputs. Varonis identified a one-click attack using the 'rovoChatPrompt' URL parameter to preload malicious instructions, which Atlassian patched server-side on July 8, 2026. PromptArmor demonstrated an indirect prompt injection via uploaded files that causes Rovo to send sensitive data to attacker-controlled servers, a vector that remains unpatched as of August 5, 2026, with no CVE assigned.

2mo ago · hacker-news

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) added CVE-2026-8037, a critical command injection vulnerability in Progress Kemp LoadMaster, to its Known Exploited Vulnerabilities (KEV) catalog due to active exploitation. The flaw allows unauthenticated attackers to execute arbitrary commands on affected devices via unsanitized input in multiple command endpoints. A total of 792 exploitation attempts have been observed from 65 unique IP addresses across 18 countries, with recent activity detected as of August 4, 2026. Federal agencies are urged to patch by August 10, 2026, per BOD 26-04.

2mo ago · hacker-news

N-able has released Hotfix 2 for its N-central Remote Monitoring and Management (RMM) product to address ongoing exploitation of a critical authentication bypass vulnerability, CVE-2026-18577, which has been actively exploited in the wild. The vulnerability, impacting versions prior to 2026.3.1.7, allows attackers to achieve remote administrative access and perform account takeover. Once inside, threat actors have used the Take Control feature to access managed systems and establish persistence via Cloudflare Tunnel services, even after N-central access was revoked.

2mo ago · hacker-news

A zero-day vulnerability in Metabase versions 1.58 and above is being actively exploited in the wild, allowing unauthenticated attackers to gain administrator access by injecting arbitrary SQL into the application database. The attack chain involves sending a POST request to '/api/session/reset_password' followed by a GET to '/api/user/current', which can be detected in logs as an indicator of compromise. Metabase Cloud instances have been patched, but self-hosted users are urged to update immediately. The PC maker Framework confirmed customer data was accessed, including names, IPs, addresses, phone numbers, and emails, though no payment data was involved.

2mo ago · hacker-news

UNC6671, a financially motivated threat actor group, is conducting vishing attacks to steal credentials and multi-factor authentication tokens by impersonating IT help desk personnel and contacting employees on their personal mobile devices. The attackers use adversary-in-the-middle (AitM) infrastructure to capture credentials and session tokens, enabling access to SaaS platforms such as Microsoft 365 and Okta. The group operates under multiple extortion brands including Redact, Pink, Helix, and Falcon, and has exfiltrated data from organizations in North America, Australia, and the U.K., collecting over $10.6 million in Bitcoin between January and May 2026. Google and CrowdStrike assess that the group leverages social engineering rather than technical vulnerabilities, highlighting the need for phishing-resistant MFA and improved session controls.

2mo ago · hacker-news

A large-scale npm supply chain campaign has distributed nearly 800 malicious packages designed to deliver a cross-platform RAT and infostealer. The packages bypass typical lifecycle hook detection by instructing developers to load them via require(), triggering a downloader named WEL1DROPPER that fetches payloads from Cloudflare Workers or fallback domains using DNS TXT record exfiltration. The payloads target Windows, Mac, and Linux systems, establishing persistence, evading detection, and deploying secondary malware such as Sliver C2. The campaign, tracked as Flooding Dropper, may target Russian financial institutions and appears to evolve from the earlier Moika campaign.

2mo ago · hacker-news

ClickFix-style attacks are delivering a Go-based macOS stealer that profiles the system, escalates privileges via a fake system error prompt, and steals sensitive data including browser passwords, Apple iCloud Keychain, and cryptocurrency wallet contents. The malware includes a 'DRAIN' routine that siphons partial or full balances from wallets supporting Bitcoin, Litecoin, Dogecoin, Monero, Ethereum, and XRP into attacker-controlled accounts. The infrastructure used in the attack is linked to Aeza Group, a Russian bulletproof hosting provider under international sanctions.

2mo ago · hacker-news

A pre-authentication reflected cross-site scripting (XSS) vulnerability in WordPress, tracked as CVE-2026-64638, affects all versions and can be exploited without authentication. The XSS flaw exists in the login screen, where a malicious username can bypass sanitization and execute JavaScript on the failed-login error page. This XSS can be chained with the SOME technique to achieve PHP code execution on the server when an administrator interacts with an attacker-controlled page, enabling actions such as plugin installation or arbitrary ZIP upload without requiring the plugin to be activated.

2mo ago · hacker-news

A critical use-after-free vulnerability in Linux's SCTP implementation, tracked as CVE-2026-64564 and named SCTPhantom, has existed since 2008 and could allow local attackers to gain root privileges and escape containers. The flaw arises from improper handling of SCTP dynamic address reconfiguration, where a delete request is validated against one address but applied to another, leading to a use-after-free condition. Tencent's Zhuque Lab demonstrated successful exploitation on multiple Linux distributions, achieving host-level root access without requiring CAP_NET_ADMIN or CAP_SYS_ADMIN under specific conditions. The vulnerability was patched in Linux kernel versions 7.1.6, 6.18.42, 6.12.101, and 6.6.148, released on August 3, 2026.

2mo ago · hacker-news

TeamPCP, a threat actor active since at least 2020, has evolved from exploiting exposed Redis, Docker, Ray, and React infrastructure to conducting large-scale supply chain attacks. The group has used overlapping infrastructure and tradecraft across campaigns, including ShadowRay 2.0 (aka IronErn) and TA-NATALSTATUS, which targeted Redis servers to deploy cryptocurrency miners. More recently, TeamPCP has poisoned open-source libraries via GitHub Actions abuse and token theft, while also deploying destructive malware such as 'kube.py' that includes wiper functionality targeting Kubernetes clusters, particularly those in Iran.

2mo ago · hacker-news

Researchers at Novee Security identified critical vulnerabilities in Anthropic's Claude Code and Google's Gemini CLI that could allow unprivileged attackers to access CI workflow secrets or execute code on CI runners. CVE-2026-12537 in Gemini CLI enables OS command injection via a malicious .gemini/.env file, allowing pre-sandbox code execution on CI hosts. CVE-2026-54316 in Claude Code allows exfiltration of API keys one character at a time through Hugging Face's public download counter. Both vulnerabilities have been patched, with no evidence of active exploitation in the wild. The root cause across both systems was flawed 'harness' code that failed to properly validate or sandbox model-generated commands.

2mo ago · hacker-news

Researcher Dirk-jan Mollema demonstrated a technique where malware running in a signed-in Windows session can abuse Windows Hello for Business keys to silently authenticate to Microsoft Entra ID, enabling persistent cloud access without extracting private keys or requiring administrator privileges. By leveraging WebAuthn, the attacker can request a signed assertion from the compromised endpoint and use it to obtain a Primary Refresh Token (PRT), register a new device, and bypass phishing-resistant authentication requirements. The technique exploits legitimate Windows ticketing behavior and does not require device-specific access, allowing attackers to establish long-term access to cloud resources under certain tenant policies.

2mo ago · hacker-news

Security researcher Malcolm Stagg disclosed a new attack class called NatJack that exploits weaknesses in NAT implementations to hijack TCP sessions, spoof DNS responses, expose mapped ports, and exhaust NAT tables. The attacks affect both Windows and Linux systems, with two specific vulnerabilities assigned CVEs: CVE-2026-56181 in Windows NAT used by Hyper-V and CVE-2026-63913 in Linux Netfilter conntrack. The technique requires an attacker with privileged access behind the same NAT as the victim and manipulates connection-tracking state to redirect or interfere with network traffic, highlighting a fundamental assumption flaw in NAT designs.

2mo ago · hacker-news

PortSwigger researchers, led by James Kettle, used an AI-assisted system called HTTP Terminator to discover novel HTTP desynchronization techniques, including a new 'dangling-byte' method that improves the reliability of response queue poisoning (RQP) attacks. The system autonomously generated and validated 30,000 candidate attack vectors, identifying over 700 vulnerable targets across banks, government systems, and critical infrastructure. A human-guided follow-up revealed a zero-day in Apache Traffic Server, later patched and assigned CVE-2026-63078, while another concept, Shared-Parser Confusion, was proposed by the AI and validated by researchers.

2mo ago · hacker-news

An active, widespread phishing campaign is targeting Microsoft 365 users using adversary-in-the-middle (AitM) techniques to hijack accounts and collect payroll and finance-related emails. The attack chain begins with voicemail-themed phishing emails that redirect through legitimate services like Google and Amazon S3 to mask malicious infrastructure. The AitM pages capture credentials and MFA codes, while JavaScript fingerprints the victim's browser and sends data to a PHP endpoint. Attackers use residential proxies to maintain long-lived, geographically proxied sessions and abuse Microsoft Graph API to enumerate and collect sensitive mailbox data, avoiding typical BEC behaviors to evade detection.

2mo ago · hacker-news

Multiple active threat campaigns were reported, including a new SideWinder attack chain using ClickOnce files to deploy Rust-based backdoors, a large-scale npm supply chain attack named 'Flooding Dropper' involving 846 malicious packages, and a Chinese threat actor leveraging a DeepSeek AI agent in an LLM-managed campaign for proxyjacking. A new XCSSET macOS malware variant (v40) spreads via compromised Xcode projects and includes a Telegram trojanizer. The Gentlemen ransomware affiliate deployed EtherRAT, which retrieves C2 data from an Ethereum smart contract. Additionally, Interlock ransomware abused Volatility3 to extract credentials from memory, and a critical RCE flaw in the Odysseus AI workspace allowed authenticated users to execute OS commands. Several phishing campaigns used fake Bank of America and Coldcard wallet lures to install ScreenConnect, while AI-powered scam farms like FunFoneFarm lower the barrier to entry for cybercrime.

2mo ago · hacker-news

Researchers from MIT CSAIL discovered a new side-channel attack technique called INTERRUPT INJECTION that exploits a timing vulnerability in Spectre v2 mitigations on Intel and AMD CPUs. By injecting hardware interrupts at a precise moment between branch predictor sanitization and kernel use, an unprivileged local program can re-poison the predictor and leak kernel memory, including sensitive data like /etc/shadow. The attack bypasses existing Safe-RET protections on AMD Zen 1 through Zen 4 processors and has been demonstrated on Zen 2 with 91.97% accuracy. A patch has been merged into the Linux kernel, but no CVE has been assigned, and Intel considers mitigation unnecessary despite demonstrated mispredictions.

2mo ago · hacker-news

Cisco has released security updates to address 12 critical vulnerabilities in its Catalyst SD-WAN and IOS XE Software, including three with CVSS scores of 9.8 or higher. The flaws involve improper input validation, access control, command injection, and other memory and logic vulnerabilities that could allow remote attackers to execute arbitrary code, escalate privileges, or access sensitive data. Additionally, Cisco patched a high-severity flaw in the Integrated Management Controller (IMC) web interface, CVE-2026-20200, for which a proof-of-concept exploit exists, allowing authenticated attackers with low privileges to achieve root-level command execution and deeply compromise server hardware trust.

2mo ago · hacker-news

A new Linux kernel vulnerability in KVM's shadow memory management unit, tracked as CVE-2026-64561 and dubbed 'Zapscape', enables a privileged attacker within an L1 guest VM to escape to the host system when nested virtualization is exposed. The flaw stems from a stale-root check ordering issue leading to a use-after-free condition during page fault handling, allowing post-free writes and potential host code execution. A public proof-of-concept demonstrates the ability to create a file on the host with root privileges, though the exploit requires adaptation for real-world use. The vulnerability affects Linux versions from 5.9 until fixed versions, with upstream patches merged and assigned CVE-2026-64561.

2mo ago · hacker-news

Maksim Silnikau, a Belarusian national, was sentenced to 16 years in prison for creating and operating Ransom Cartel, a ransomware-as-a-service (RaaS) platform active from 2021 to 2023. The operation facilitated attacks on at least 18 organizations globally, leveraging stolen credentials and a hidden affiliate panel for managing ransom negotiations and payments via cryptocurrency mixers. While Silnikau did not conduct most intrusions himself, he built and managed the infrastructure, including a ratings system for affiliates. The operation was initially launched under a different name in May 2021 before being rebranded as Ransom Cartel later that year.

2mo ago · hacker-news

A new threat technique called AI Recommendation Poisoning is being used by companies to manipulate AI assistant behavior by injecting memory-altering instructions through 'Ask AI' buttons on websites. When users click these buttons, pre-filled deep links execute prompts that silently mark specific domains as trusted sources in the AI's long-term memory, biasing future responses without user consent. This technique leverages legitimate AI features and bypasses traditional content-based defenses, with evidence of widespread adoption across marketing tools and CMS plugins. The attack persists indefinitely in the AI's memory and can influence security and product evaluation decisions.

2mo ago · hacker-news

Researchers Talal Haj Bakry and Tommy Mysk discovered a privacy issue in Apple's iCloud Private Relay that can expose a user's real IP address due to WebKit proxy bypasses. The vulnerability stems from three WebKit features—DNS prefetching, WebAuthn Related Origin Requests, and WebTransport—which send traffic directly from the device, bypassing the proxy configuration. This allows websites, particularly those using WebAuthn for passkeys, to view a user's real IP address even when Private Relay is enabled, without requiring user interaction or passkey usage.

2mo ago · hacker-news

A weak random number generator in the CryptoJS library, specifically CryptoJS.lib.WordArray.random(), was exploited to compromise cryptocurrency wallet recovery phrases, leading to the theft of approximately $5.7 million across two attack waves. The vulnerability affected five wallet applications—RRWallet, Bexo Wallet, NanChat, Bitcoin Libre, and Milo—where weak entropy reduced the effective search space for recovery phrases from 2^128 or 2^256 down to roughly 2^39 or 2^47, making them brute-forceable. The flaw was exploited in attacks between May and July 2026, with stolen funds traced across Bitcoin, Ethereum, Tron, Rootstock, and Polygon blockchains. The issue stems from a regression in CryptoJS versions prior to 4.0.0, where a previously fixed issue was reintroduced in version 3.3.0 due to backward compatibility concerns.

2mo ago · hacker-news

CISA has added CVE-2026-63077, a critical remote code execution vulnerability in JetBrains TeamCity, to its Known Exploited Vulnerabilities catalog due to active exploitation in the wild. The flaw stems from deserialization of untrusted data in the TeamCity agent polling protocol, allowing unauthenticated attackers to bypass authentication and execute arbitrary commands with the privileges of the TeamCity server process. A successful exploit could lead to theft of sensitive data, configuration, and credentials, as well as compromise of build artifacts and CI/CD pipelines. Federal agencies are required to patch by August 8, 2026, per BOD 26-04.

2mo ago · hacker-news

Zbtlink routers are shipped with a factory-implanted backdoor named ENDLESSDOORS, which establishes unauthenticated root shells by connecting to command-and-control servers. The backdoor, based on the 'rctl' tool, runs as a disguised kernel thread with root privileges and contacts C2 infrastructure every 35 seconds. It allows remote attackers to execute arbitrary commands or spawn interactive root shells without authentication, enabling full device takeover. The backdoor is present in at least 20 router models, all of which initiate connections to a shared set of C2 endpoints.

2mo ago · hacker-news

Security researchers discovered critical flaws in agent infrastructures from AWS, Google, and Vercel that allow attackers to bypass model authorization and directly trigger tool execution. The vulnerabilities, collectively named CoreBreak, stem from insufficient validation of tool-call inputs, enabling untrusted or forged instructions to reach execution without model oversight. AWS addressed CVE-2026-18830 in its managed Bedrock AgentCore service by adding server-side validation, while Google patched two separate issues in its Agent Development Kit (ADK) for Python, including CVE-2026-18236 for continuation forgery and a resumable-mode bypass. Vercel fixed two related authorization bypasses in its AI SDK harness packages, tracked as CVE-2026-64650 and CVE-2026-64651, which allowed sandboxed malicious code to invoke host tools without model authorization.

2mo ago · hacker-news

Attackers exploited a SQL injection vulnerability in a public-facing web application to gain access to an Oracle database, then leveraged Oracle's embedded Java Virtual Machine to compile and execute malicious Java code within the database itself. This post-exploitation toolkit, tracked as khunt by Huntress, allowed the attackers to achieve SYSTEM-level privilege escalation on the underlying Windows server without writing files to disk. The toolkit consists of multiple Java classes and PL/SQL wrappers that enable command execution, credential theft, file system access, and registry hive extraction. The malicious activity was traced to a single IP address, and detection requires hunting for specific object names and SQL patterns due to the in-memory nature of the attack.

2mo ago · hacker-news

Connor Riley Moucka pleaded guilty to orchestrating breaches of Snowflake customer accounts in 2024, exploiting stolen credentials from infostealer malware that had not been rotated and where multi-factor authentication (MFA) was disabled. The attacks impacted at least 165 organizations and exposed data belonging to over 100 million people, including sensitive personal and government-related information. No platform vulnerability was exploited; instead, the campaign leveraged previously compromised credentials, with Mandiant attributing the activity to threat actor UNC5537. Moucka monetized the stolen data through ransoms and sales, netting at least $495,000, while re-extorting at least one victim using data of a government official and their family.