Google has adjusted its Chrome bug bounty program by increasing rewards for complex memory safety vulnerabilities after previously reducing payouts. The changes focus on incentivizing reports of memory safety flaws with higher base rewards and specific bonuses for fuzzing innovations. Non-memory safety issues receive moderate rewards if they lead to fixes. This restructuring aims to better align incentives with impactful vulnerability discovery amid evolving threat landscapes influenced by AI.
FreeRDP versions before 3.32.0 contain a vulnerability in the freerdp_http_request() function where TLS handshakes are completed without proper server certificate chain validation. This flaw allows on-path attackers to impersonate login.microsoftonline.com during the Azure AD authentication process, potentially capturing OAuth authorization codes and substituting the OpenID discovery document. The vulnerability has a CVSS score of 6.8, indicating a high severity level.
Criminal IP by AI SPERA introduces AITEM, an AI-powered evolution of Attack Surface Management (ASM) that extends beyond traditional asset discovery. AITEM integrates threat intelligence with AI to connect exposure discovery, investigation, prioritization, and automated response. It aggregates data from external assets, internal infrastructure, open-source intelligence, dark web sources, leaked data, and emerging vulnerabilities to provide a comprehensive threat exposure management approach. The solution aims to help security teams focus on meaningful risks and accelerate response times amid increasing attacker automation. This announcement reflects a broader industry trend toward AI-driven, integrated security operations.
In July 2026, a sophisticated malware campaign attributed to the TraderTraitor group was uncovered, leveraging a trojanized Terraform provider to target cryptocurrency and Web3 developers. The attack uses a malicious Terraform provider that downloads a Bash loader from a HashiCorp-themed lookalike domain, which then deploys platform-specific payloads for macOS, Linux, and Windows. Encrypted executables are hidden in fake font files and decrypted with AES-256-CBC. The campaign delivers FLATROOF, a Rust-based backdoor with multiple command and control (C2) channels, and Python-based information stealers targeting browser credentials and cryptocurrency wallet data. The attack concludes with deployment of the ROOFDECK backdoor, which uses resilient C2 discovery methods including cryptographically signed Pastebin dead drops and Nostr metadata.
This content discusses how frontier AI models are transforming vulnerability management, particularly for Microsoft software. AI accelerates vulnerability discovery and patch development, resulting in a significant increase in identified vulnerabilities, especially for on-premises Microsoft products. Microsoft mitigates most cloud vulnerabilities without customer action, but on-premises customers should expect more frequent and numerous patches. CISOs face challenges balancing rapid patching with accuracy at unprecedented scale. The post recommends increased patching resources, faster patch deployment for critical systems, use of AI-powered scanning harnesses, and defense-in-depth strategies. Microsoft also collaborates with industry peers to address open-source vulnerabilities and promotes Secure by Design and Secure by Default principles to improve security posture.
Remote Monitoring and Management (RMM) software used by Managed Service Providers (MSPs) grants privileged access across multiple customer environments, making its security controls critical to limit risk. This report outlines eight essential security controls MSPs should test when evaluating RMM platforms, including endpoint discovery, patch management, privileged access controls, alert prioritization, secure automation, integration with security operations, recovery readiness, and tenant separation. Recent incidents involving RMM vulnerabilities and exploitation highlight the importance of these controls. MSPs must ensure that RMM tools provide strong identity management, tenant isolation, and seamless integration with security and recovery workflows to reduce the attack surface and potential blast radius from compromised accounts or endpoints.
Google has temporarily paused accepting product vulnerability reports through its Open Source Software Vulnerability Reward Program (OSS VRP) due to a surge in invalid automated submissions. This pause affects only product vulnerability reports and does not impact supply chain reports or pending submissions. Google encourages researchers to submit vulnerabilities through other programs such as the Cloud VRP or Patch Rewards Program. The company plans to revisit and update the OSS VRP in early 2027. This action follows similar adjustments in Google's Chrome and Android bug bounty programs in response to increased AI-assisted vulnerability discovery.
This report summarizes multiple cybersecurity developments including a set of iCloud email spoofing vulnerabilities, phishing campaigns targeting AI policy experts, a Chrome adblocker extension spying on AI chat conversations, and a large batch of vulnerabilities patched by Kiteworks. The iCloud spoofing bugs allowed attackers to send emails from arbitrary icloud.com addresses that passed SPF, DKIM, and DMARC checks due to inconsistent mail header parsing. The phishing campaigns involved a China-aligned group using fake AI advisory committee invitations to steal credentials via an adversary-in-the-middle proxy. The Poper Blocker Chrome extension was found to collect users' browsing history and AI chat conversations stealthily. Kiteworks patched over 100 vulnerabilities, including critical bugs that could lead to account takeover and code execution. Additional highlights include AI-assisted discovery of Android app flaws and a Cloudflare Containers data leakage vulnerability that was patched. No active exploits in the wild were reported for these issues.
RemusStealer is a malware-as-a-service program exhibiting multiple similarities to LummaStealer, including anti-virtual machine checks, credential theft tactics, and Application-Bound Encryption bypass methods. A key distinction is RemusStealer's use of Ethereum smart contracts for C2 communications instead of Steam or Telegram dead drop resolvers. Written in Go, the stealer performs extensive system discovery, gathering information about antivirus products, hardware specifications, and operating systems. It employs hidden desktop environments to evade detection while launching Microsoft Edge and Brave browser processes on non-primary window stations to steal credential files. The malware establishes connections to suspicious domains ending in .shop or .biz and communicates with Ethereum-related infrastructure. This evolution demonstrates attackers' increasing sophistication in leveraging blockchain technology to masquerade malicious communications through smart contracts.
Microsoft says cyberattackers are currently benefiting from artificial intelligence faster than defenders, allowing threat actors to speed up vulnerability discovery, malware development, and post-compromise activity while security teams struggle to keep pace. [...]
Google Threat Intelligence Group (GTIG) reports that AI is significantly accelerating the discovery and exploitation of software vulnerabilities, with AI-discovered flaws more likely to enable remote code execution. The volume of disclosed vulnerabilities doubled in 2026, with a notable increase in high-risk and exploited vulnerabilities. AI-discovered vulnerabilities tend to have higher impact, with 50% enabling remote code execution compared to 26% for non-AI discoveries. Exploitation of known vulnerabilities (n-days) is rising, possibly due to AI-assisted automation in weaponizing these flaws. AI-related vulnerabilities in AI systems themselves are also increasing, though zero-day exploitation in AI infrastructure has not yet been observed. The trend suggests continued growth in vulnerability discovery and exploitation driven by AI capabilities.
Improper Verification of Source of a Communication Channel in the ADS discovery of the Go implementation of Apache PLC4X (PLC4Go) allows an attacker able to send UDP datagrams to the discovering host to redirect subsequent connections to an arbitrary, attacker-chosen address. The discovery result's connection
address was derived from the AmsNetId claimed in the response body rather than from the datagram's actual source address. One spoofed discovery response can therefore insert an inventory entry pointing at any host, including hosts outside the local network, and an application that connects to discovered devices
will open its ADS session, including any configured route credentials, to that host.
Additionally, discovery listeners in both implementations can be disabled by a single malformed datagram:
- In PLC4Go ADS discovery, a short version block causes a panic that ends the listener for the rest of the discovery call, so legitimate devices answering afterwards are not reported.
- In PLC4J, the ADS and EtherNet/IP discoverers stop on an unhandled exception from a malformed response.
- The PLC4J Modbus discoverer can be made to spin indefinitely, consuming a CPU core, by a scanned host that sends a partial response.
Exploitation requires the application to invoke the discovery API, which is opt-in, and for the connection redirect, to act on the discovered items.
This issue affects Apache PLC4X: PLC4Go from 0.11.0 before 1.0.0; PLC4J ADS and Modbus drivers from 0.10.0 before 1.0.0; PLC4J EtherNet/IP driver from 0.11.0 before 1.0.0. PLC4Go is consumed as the Go module github.com/apache/plc4x/plc4go; versions refer to the corresponding Apache PLC4X releases.
Users are recommended to upgrade to version 1.0.0, which fixes the issue. Version 1.0.0 derives the connection address from the datagram's source address and logs a warning when the claimed AmsNetId disagrees with it.
Improper Verification of Cryptographic Signature and Improper Certificate Validation in the OPC UA driver of Apache PLC4X (PLC4J) allows an attacker in a network position between client and server to impersonate the OPC UA server and to read, forge or modify secure-channel traffic, including user credential ssent by the client.
The defect manifests differently depending on the version:
- In 0.9.0 through 0.11.0 a failed message-signature check is only logged and never enforced, and there is no mechanism to verify the server certificate: it is taken from the unauthenticated GetEndpoints discovery response and used to encrypt the user's password.
- In 0.12.0 through 0.13.1 the signature check is inverted (valid signatures are rejected, invalid ones accepted), and server certificates are accepted without a trust anchor by default.
- In all affected versions the default security policy is None. Starting with 0.12.0 the driver additionally continues silently at a weaker security policy than the one configured, and starting with 0.13.0 endpoint selection prefers the weakest matching endpoint.
Users checking only for one of these mechanisms may wrongly conclude they are unaffected.
This issue affects Apache PLC4X: from 0.9.0 before 1.0.0.
Users are recommended to upgrade to version 1.0.0, which fixes the issue. Version 1.0.0 verifies message signatures correctly, refuses to connect unless the server certificate can be verified against a configured trust store or pinned certificate, defaults to Basic256Sha256 with SignAndEncrypt, and fails the
connection if the negotiated security policy is weaker than the configured one.
Times Car, a Japanese car-sharing service, confirmed a data breach affecting approximately 6.6 million current and former user accounts. The breach involved unauthorized access to personal information including names, addresses, dates of birth, contact details, driver’s license data, identity verification documents, account passwords (stored in a non-reversible form), and linked service IDs. Credit card information was not compromised. The company blocked access shortly after discovery and is conducting a forensic investigation with external experts. Notifications to affected users are being sent in stages. There is currently no evidence that the stolen data has been distributed online.
This report summarizes multiple cybersecurity incidents and emerging threats including the takeover of the Clop ransomware gang's leak site by the ShinyHunters group, the BragJack vulnerability that allows malicious browser extensions to hijack AI assistants, a new Go implant malware targeting AI agent tooling, and a Docker botnet that steals AI API keys. Additionally, it covers exposed remote access credentials in the US water utility sector, a novel Windows implant using commercial AI models for command decisions, and a high-severity pre-authentication denial-of-service vulnerability in TDengine industrial telemetry software. The report also notes Canonical's changes to Ubuntu kernel update cycles to accelerate patching and the discovery of a new Android banking trojan with AI-assisted development. These events highlight evolving attack methods leveraging AI and targeting critical infrastructure and cloud-native environments.
A vulnerability in the Linux kernel's IPv6 implementation allows a UDP length overflow when Path MTU Discovery (PMTU) is enabled with large MTU settings. This occurs because the UDP length field can overflow its 16-bit limit when sending large packets over devices with an MTU larger than the IPv6 maximum MTU. The issue triggers a kernel warning and affects UDP sockets when IPV6_MTU_DISCOVER is set to IPV6_PMTUDISC_DO or IPV6_PMTUDISC_PROBE. The flaw has been fixed in specific Linux kernel commits.
CVE-2026-82371 is a vulnerability in Brocade SANnav discovery service where sensitive authentication data is logged in plaintext. This allows individuals with read access to log files to retrieve administrative credentials and active session tokens. The flaw affects Brocade SANnav versions prior to 3.0.1a. Exploiting this vulnerability could lead to compromise of managed network infrastructure and unauthorized access to active application sessions.
TREK is a collaborative travel planner. Prior to 3.3.0, TREK allows an authenticated user to store an attacker-controlled llm_base_url through the settings API when the LLM_PARSING feature is enabled. Write permission to the target trip instance is required to trigger the vulnerable AI-assisted import path. The value is consumed by the clients in server/src/nest/llm-parse/clients/openai-compatible.client.ts, server/src/nest/llm-parse/clients/anthropic.client.ts, and server/src/nest/llm-parse/router/ollama-format.client.ts without applying the server-side request forgery guard. Triggering AI-assisted trip parsing causes the server to request the supplied destination, and upstream error response text can be returned in parsing warnings. This permits internal service discovery and access to link-local cloud metadata, with possible disclosure of infrastructure credentials and subsequent modification of protected cloud resources. This issue is fixed in version 3.3.0.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (920277566757bdc38563c7e2150e906790478cbd7a0c48f0bb53004e82eec4ca)
The package impersonates an Apple/Unity StoreKit internal namespace (com.apple.unityplugin.storekit) and, on require(), index.js performs an HTTPS GET to a hardcoded Project Discovery Interactsh (oast.fun) subdomain, transmitting the package name, os.platform(), and os.hostname() as query parameters (target: https://dapnhid534ch06s9vpm0mbg1httu5gytc.oast.fun). The behavior is unconditional, undocumented, and fires at module load with no caller opt-in. The name/scope, the author claim of 'Apple, Inc', and the OAST callback destination are the standard dependency-confusion reconnaissance shape: the beacon confirms that an internal build system has resolved the attacker's public package instead of Apple's private one, and leaks the internal hostname of the machine that resolved it.