DNG SDK versions 1.7.1 2536 and earlier are affected by an out-of-bounds read vulnerability that could lead to disclosure of sensitive memory. An attacker could leverage this vulnerability to disclose sensitive information. Exploitation of this issue requires user interaction in that a victim must open a malicious file.
CVE-2026-39492 is a critical unauthenticated SQL Injection vulnerability in the WP Maps plugin by Flipper Code – WordPress Development Company, affecting versions up to and including 4.9.1. This vulnerability allows an attacker to inject SQL commands without authentication, potentially leading to high confidentiality impact and limited availability impact. No official patch or remediation guidance is currently available from the vendor. Users should monitor vendor advisories for updates and consider mitigating exposure until a fix is released.
Oleksii Oleksiyovych Lytvynenko admitted to working on the development of a loader for the Conti gang. The post Ukrainian Man Pleads Guilty in US to Conti Ransomware Charges appeared first on SecurityWeek .
A sophisticated malware campaign exploits growing interest in artificial intelligence by distributing malicious files disguised as AI-related learning resources and technical guides. The attack employs an exceptionally complex multi-stage infection chain beginning with compressed archives containing LNK shortcuts and hidden PDF files. Through multiple layers of obfuscation involving PowerShell scripts, batch files, and AutoHotkey loaders, the campaign establishes persistent access and deploys two distinct .NET Remote Access Trojans including AsyncRAT. The intermediate scripts extensively use Simplified Chinese variable names and exhibit coding patterns suggesting AI-assisted development, with cultural references to Chinese mythology used as symbolic aliases for Windows API calls. The attack implements advanced techniques including process hollowing, reflective DLL injection, and scheduled task persistence while actively disabling Windows Defender exclusions to facilitate execution.
Iranian and Russian shadow fleet vessels are utilizing sophisticated online infrastructure consisting of over 36 inauthentic websites to facilitate sanctions evasion. These websites impersonate ship registries, national maritime administrations, seafarer training organizations, protection and indemnity clubs, and classification societies from jurisdictions including Comoros, Benin, Bhutan, Cameroon, Chad, Equatorial Guinea, Gambia, Haiti, Malawi, Nicaragua, and Zambia. The infrastructure operates through three identified clusters designated Alpha, Bravo, and Charlie, which demonstrate technical overlaps suggesting a broader ecosystem supporting multiple sanctions evasion networks. Operators employ tactics including automated document generation, typosquatting, identity spoofing, and mutual endorsement loops between fraudulent entities. Attribution includes links to Indian web development company Oceaniek Technologies and two Syrian nationals. The infrastructure has documented connections to seventeen vesse...
OS command injection in the NodejsFunction local bundling pipeline in aws-cdk-lib before 2.245.0 (2.246.0 on Windows) might allow an actor who controls the value of one or more bundling properties (externalModules, define, loader, inject, or esbuildArgs) to execute arbitrary commands on the host running the CDK toolchain via injected shell metacharacters. This issue requires the threat actor to control the value of one or more of the affected bundling properties in the CDK application.
To remediate this issue, users should upgrade to aws-cdk-lib 2.245.0 (2.246.0 on Windows) or later.
ESF-IDF is the Espressif Internet of Things (IOT) Development Framework. In versions 5.2.6, 5.3.5, 5.4.4, 5.5.3, and 6.0, an out-of-bounds read exists in the BlueDroid AVRCP vendor-command parser (avrc_pars_vendor_cmd() in components/bt/host/bluedroid/stack/avrc/avrc_pars_tg.c). This issue has been patched in versions 5.2.7, 5.3.6, 5.4.5, 5.5.4, and 6.0.1.
ESF-IDF is the Espressif Internet of Things (IOT) Development Framework. In versions 5.2.6, 5.3.5, 5.4.4, 5.5.4, and 6.0, a heap buffer overflow exists in the Security Scheme 2 (SRP6a) session-setup path of the protocomm component. The first-phase handler (handle_session_command0() in components/protocomm/src/security/security2.c) trusts the length of a client-supplied protobuf field for the SRP6a username and copies it into a buffer whose size is derived from a narrower destination type. The resulting truncation-versus-copy asymmetry corrupts the heap when an oversized value is supplied. This issue has been patched in versions 5.2.7, 5.3.6, 5.4.5, 5.5.5, and 6.0.1.
ESF-IDF is the Espressif Internet of Things (IOT) Development Framework. In versions 5.5.4 and 6.0, several ESP-TEE secure-service wrappers in esp_secure_services.c and esp_secure_services_iram.c validated only some of the caller-supplied pointer arguments, leaving input pointer arguments unchecked. Because the underlying TEE-protected hardware peripherals (e.g., ECC, SHA, SPI) run in RISC-V machine mode (M-mode) with full address-space access, a caller could supply pointers into TEE-exclusive memory as inputs, causing the peripheral to read TEE memory and return results derived from it to the REE. Depending on the wrapper, the result contains raw bytes from TEE memory, a computed function of TEE memory recoverable through repeated calls, or a single bit per call that forms an oracle for incremental disclosure of TEE-resident sensitive data. This issue has been patched in versions 5.5.5 and 6.0.1.
CVE-2026-45328 is a critical vulnerability in the Espressif IoT Development Framework (esp-idf) versions 5.5.4 and 6.0. It involves improper input validation in the esp_tee component, which exposes secure-service wrappers bridging calls from user applications to TEE-protected hardware peripherals and security features. This vulnerability has been patched in versions 5.5.5 and 6.0.1.
ESF-IDF is the Espressif Internet of Things (IOT) Development Framework. In versions 5.2.7, 5.3.5, 5.4.4, 5.5.4, and 6.0.1, an out-of-bounds read flaw exists in the DHCP server option parser (parse_options() in components/lwip/apps/dhcpserver/dhcpserver.c) shipped with ESP-IDF's lwIP component. The parser walks the BOOTP/DHCP options field without validating that each option's length byte and declared payload length stay within the received packet buffer. A crafted DHCP request can cause the parser to read past the end of the options buffer into adjacent heap memory. The issue affects the DHCP server used by ESP-IDF's SoftAP and any configuration where the device runs as a DHCP server on a local network. This issue has been patched in versions 5.2.8, 5.3.6, 5.4.5, 5.5.5, and 6.0.2.
ESF-IDF is the Espressif Internet of Things (IOT) Development Framework. In versions 5.2.6, 5.3.5, 5.4.4, 5.5.4, and 6.0, a NULL-pointer dereference exists in the WebSocket subprotocol-negotiation path of the esp_http_server component. While parsing the client-supplied Sec-WebSocket-Protocol request header during the WebSocket handshake, the tokenisation result is dereferenced without a NULL check, so a malformed header value can crash the server before any application-level authentication runs. This issue has been patched in versions 5.2.7, 5.3.6, 5.4.5, 5.5.5, and 6.0.1.
Pipecat versions from 0.0.90 up to but not including 1.2.0 contain a path traversal vulnerability in the development runner component. This flaw allows an unauthenticated attacker with network access to read arbitrary files accessible by the pipecat process by exploiting improper handling of URL-encoded path parameters. The vulnerability is fixed in version 1.2.0.
Anthropic's Claude Mythos Preview, a public large language model (LLM) with safeguards disabled, can rapidly generate working exploits targeting known vulnerabilities within hours or even minutes. It has demonstrated the ability to create proof-of-concept exploits for Firefox and Windows vulnerabilities, significantly accelerating the exploit development process for N-day vulnerabilities. This reduces the traditional patch gap risk, as exploits can be weaponized before most systems receive patches. The cost and expertise barriers to exploit development are lowered, expanding the pool of potential attackers. This shift calls for updated patching strategies emphasizing speed and urgency.
Atsign has introduced AI Architect, a platform that applies cryptographic invisibility to agentic AI-built applications. This approach secures application identities by making them effectively invisible to attackers, thereby preventing exploitation of vulnerabilities through credential or identity attacks. The platform assigns unique cryptographic identities to all resources involved in app development, ensuring that credentials are never exposed in cleartext and that no open ports or public APIs are available for scanning. AI Architect integrates security by design into AI coding workflows, producing precise prompts that guide secure code generation and governance. This method aims to neutralize vulnerabilities by protecting the primary attack vector—identity—rather than eliminating the vulnerabilities themselves.
In the Linux kernel, the following vulnerability has been resolved: hfsplus: fix held lock freed on hfsplus_fill_super() hfsplus_fill_super() calls hfs_find_init() to initialize a search structure, which acquires tree->tree_lock. If the subsequent call to hfsplus_cat_build_key() fails, the function jumps to the out_put_root error label without releasing the lock. The later cleanup path then frees the tree data structure with the lock still held, triggering a held lock freed warning. Fix this by adding the missing hfs_find_exit(&fd) call before jumping to the out_put_root error label. This ensures that tree->tree_lock is properly released on the error path. The bug was originally detected on v6.13-rc1 using an experimental static analysis tool we are developing, and we have verified that the issue persists in the latest mainline kernel. The tool is specifically designed to detect memory management issues. It is currently under active development and not yet publicly available. We confirmed the bug by runtime testing under QEMU with x86_64 defconfig, lockdep enabled, and CONFIG_HFSPLUS_FS=y. To trigger the error path, we used GDB to dynamically shrink the max_unistr_len parameter to 1 before hfsplus_asc2uni() is called. This forces hfsplus_asc2uni() to naturally return -ENAMETOOLONG, which propagates to hfsplus_cat_build_key() and exercises the faulty error path. The following warning was observed during mount: ========================= WARNING: held lock freed! 7.0.0-rc3-00016-gb4f0dd314b39 #4 Not tainted ------------------------- mount/174 is freeing memory ffff888103f92000-ffff888103f92fff, with a lock still held there! ffff888103f920b0 (&tree->tree_lock){+.+.}-{4:4}, at: hfsplus_find_init+0x154/0x1e0 2 locks held by mount/174: #0: ffff888103f960e0 (&type->s_umount_key#42/1){+.+.}-{4:4}, at: alloc_super.constprop.0+0x167/0xa40 #1: ffff888103f920b0 (&tree->tree_lock){+.+.}-{4:4}, at: hfsplus_find_init+0x154/0x1e0 stack backtrace: CPU: 2 UID: 0 PID: 174 Comm: mount Not tainted 7.0.0-rc3-00016-gb4f0dd314b39 #4 PREEMPT(lazy) Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.15.0-1 04/01/2014 Call Trace: <TASK> dump_stack_lvl+0x82/0xd0 debug_check_no_locks_freed+0x13a/0x180 kfree+0x16b/0x510 ? hfsplus_fill_super+0xcb4/0x18a0 hfsplus_fill_super+0xcb4/0x18a0 ? __pfx_hfsplus_fill_super+0x10/0x10 ? srso_return_thunk+0x5/0x5f ? bdev_open+0x65f/0xc30 ? srso_return_thunk+0x5/0x5f ? pointer+0x4ce/0xbf0 ? trace_contention_end+0x11c/0x150 ? __pfx_pointer+0x10/0x10 ? srso_return_thunk+0x5/0x5f ? bdev_open+0x79b/0xc30 ? srso_return_thunk+0x5/0x5f ? srso_return_thunk+0x5/0x5f ? vsnprintf+0x6da/0x1270 ? srso_return_thunk+0x5/0x5f ? __mutex_unlock_slowpath+0x157/0x740 ? __pfx_vsnprintf+0x10/0x10 ? srso_return_thunk+0x5/0x5f ? srso_return_thunk+0x5/0x5f ? mark_held_locks+0x49/0x80 ? srso_return_thunk+0x5/0x5f ? srso_return_thunk+0x5/0x5f ? irqentry_exit+0x17b/0x5e0 ? trace_irq_disable.constprop.0+0x116/0x150 ? __pfx_hfsplus_fill_super+0x10/0x10 ? __pfx_hfsplus_fill_super+0x10/0x10 get_tree_bdev_flags+0x302/0x580 ? __pfx_get_tree_bdev_flags+0x10/0x10 ? vfs_parse_fs_qstr+0x129/0x1a0 ? __pfx_vfs_parse_fs_qstr+0x3/0x10 vfs_get_tree+0x89/0x320 fc_mount+0x10/0x1d0 path_mount+0x5c5/0x21c0 ? __pfx_path_mount+0x10/0x10 ? trace_irq_enable.constprop.0+0x116/0x150 ? trace_irq_enable.constprop.0+0x116/0x150 ? srso_return_thunk+0x5/0x5f ? srso_return_thunk+0x5/0x5f ? kmem_cache_free+0x307/0x540 ? user_path_at+0x51/0x60 ? __x64_sys_mount+0x212/0x280 ? srso_return_thunk+0x5/0x5f __x64_sys_mount+0x212/0x280 ? __pfx___x64_sys_mount+0x10/0x10 ? srso_return_thunk+0x5/0x5f ? trace_irq_enable.constprop.0+0x116/0x150 ? srso_return_thunk+0x5/0x5f do_syscall_64+0x111/0x680 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0033:0x7ffacad55eae Code: 48 8b 0d 85 1f 0f 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 49 89 ca b8 a5 00 00 8 RSP: 002b ---truncated---
AI-driven development is not something organizations can or should block. But it must be governed. The post Everybody Is Vibe Coding But Nobody Told the Security Team appeared first on SecurityWeek .
Anthropic has proposed industry-wide coordination to enable a temporary pause or slowdown in advanced AI development if risks escalate. The concern is that rapid AI improvements, including potential recursive self-improvement, could lead to humans losing control over AI systems. The proposal includes mechanisms for verifying that global AI labs comply with any agreed slowdown to prevent bad actors from secretly advancing. This call for coordination aims to manage AI safety risks as AI capabilities accelerate. There is no direct vulnerability or exploit currently reported, and the severity is assessed as low based on the available information.
A recently highlighted security concern involves malicious code embedded in configuration files within cloned repositories. The Miasma worm demonstrates how a single commit can wire a malicious dropper into multiple configuration files across popular development tools and package managers, such as VS Code, npm, and Composer. This means that simply cloning and opening a repository can execute harmful code without needing to install any malicious dependencies. The issue is exacerbated by the fact that these configuration files are often not reviewed during pull requests, increasing the risk of unnoticed malicious code execution.
Scope: AWS Content Type: Important (requires attention) Publication Date: 2025/07/23 6:00 PM PDT Updated Date: 2025/07/25 6:00 PM PDT Description: Amazon Q Developer for Visual Studio Code (VS Code) Extension is a development tool that integrates Amazon Q's AI-powered coding assistance directly into the VS Code integrated development environment (IDE). AWS is aware of and has addressed an issue in the Amazon Q Developer for VS Code Extension, which is assigned to CVE-2025-8217. AWS Security has inspected the code and determined the malicious code was distributed with the extension but was unsuccessful in executing due to a syntax error. This prevented the malicious code from making changes to any services or customer environments. We will update this bulletin if we have additional information to share. Affected version: Amazon Q Developer for Visual Studio Code Extension (version 1.84.0)