A flaw was found in the PathResourceManager component of the Undertow web server. This component is responsible for serving static files from the file system. The issue occurs because the mechanism used to verify if a symbolic link points to a safe location only checks if the target path starts with a specific string, without ensuring it matches a full directory name. This could allow an attacker to access files in a sibling directory if that directorys name starts with the same characters as an authorized safe directory.
The webhook platform gateway in Nous Research Hermes Agent versions builds the session chat_id as webhook:{route_name}:{delivery_id} and re-derives the authorizing route by splitting that string, so a route name containing a ":" resolves toolset authorization to a different, higher-privileged prefix route than the one authenticated, which allows a remote attacker with the target HMAC to gain that higher-privileged route's toolsets.
CVE-2026-114253 is a file inclusion vulnerability in the oretnom23 file-management-system-php affecting the include function in index.php. Manipulating the 'page' argument can lead to remote file inclusion. The vulnerability has been publicly disclosed, but no patch or version details are available due to the product's rolling release model. The vendor has been informed but has not responded yet.
Tracebit has developed 'context bombs,' deceptive prompts designed to mislead attacking AI agents by presenting fake targets and complex decoys. These context bombs significantly reduce AI attack success rates—up to 91% in some cases—and increase the cost of successful attacks by up to 23 times. The technique involves planting crafted prompts that interfere with the attacker’s objectives, causing AI agents to avoid real targets or waste resources on traps. The most effective prompt bomb mimics a fake pen testing report that mislabels real attack paths as traps and decoys as valuable targets. Combining multiple context bombs yields the strongest defense, with some AI models failing entirely to achieve their objectives. This approach effectively raises the difficulty and cost of AI-driven attacks, providing defenders with more time to respond.
CVE-2026-114252 is a medium severity SQL injection vulnerability in the itsourcecode Sales-and-Inventory-System. The flaw exists in an unknown function within the /pages/cust_searchfrm.php file, where manipulation of the ID argument can lead to SQL injection. The vulnerability can be exploited remotely without user interaction. The project uses a rolling release model, so specific affected or fixed versions are not clearly identified. The issue was reported early but has not yet been addressed by the vendor. Public exploit code is available, increasing the risk of exploitation.
Rapid7's Q3 2026 threat analysis highlights that high-severity vulnerabilities are exploited rapidly, with an average of three days from disclosure and 71% exploited within the first week. Ransomware remains the dominant threat globally, particularly targeting business services, manufacturing, healthcare, and government sectors, with the US being the most targeted country. AI is involved operationally in only 8% of incidents tracked, mostly in a non-autonomous role, assisting in tactical automation rather than fully automated attacks. Nation-state actors remain active, with multiple groups linked to Russia, China, North Korea, and Pakistan targeting government and critical sectors. The dark web continues to be a marketplace for compromised credentials, network access, and exploit sales. The report warns of a potential future increase in AI-driven automated attacks but notes current attacks are still human-directed.
### Impact
An attacker who uses this vulnerability can craft a PDF which leads to arbitrary image parsers being called. This requires loading an image which claims to use the `/DCTDecode` filter (JPEG data).
Users which do not have the optional *Pillow* dependency installed in their environment are not affected.
### Patches
This has been fixed in [pypdf==6.21.1](https://github.com/py-pdf/pypdf/releases/tag/6.21.1).
### Workarounds
If you cannot upgrade yet, consider applying the changes from PR [#4222](https://github.com/py-pdf/pypdf/pull/4222).
A vulnerability has been found in itsourcecode Sales-and-Inventory-System up to 0a3c45a78aaff3601b8cf9ce8ba82ff5c345caac. Impacted is an unknown function of the file /pages/cust_del.php. Such manipulation of the argument ID leads to sql injection. The attack can be launched remotely. The exploit has been disclosed to the public and may be used. This product does not use versioning. This is why information about affected and unaffected releases are unavailable. The project was informed of the problem early through an issue report but has not responded yet.
CVE-2026-114250 is a medium-severity SQL injection vulnerability in the itsourcecode Sales-and-Inventory-System. The flaw exists in the processing of the ID argument in the /pages/cust_edit.php file, allowing remote attackers to manipulate SQL queries. The product uses continuous delivery with rolling releases, so no specific affected or fixed versions are identified. The vulnerability was reported early, but the vendor has not responded or issued a fix. Exploit code has been published, but no known in-the-wild exploitation is confirmed.
The pfSense REST API package is an unofficial, open-source REST and GraphQL API for pfSense CE and pfSense Plus firewalls. A privilege escalation vulnerability in pfSense-pkg-RESTAPI prior to version 2.9.0 allows authenticated accounts holding the dedicated settings sync permission to modify REST API configuration secrets. This enables a non-administrative user to elevate their privileges to a full administrator. pfSense-pkg-RESTAPI 2.9.0 patches the issue. Access to the HA settings sync endpoint now strictly requires full administrative privileges (`page-all`). The sub-administrative sync privilege has been removed in versions v2.9.0 and later. In addition to correcting the access control boundary, version 2.9.0 includes the following hardening measures: opt-in default and format and validation modernization. As a workaround, one may mitigate exposure by taking either or all of the following steps. Revoke Scoped Permissions: Remove the `api-v2-system-restapi-settings-sync-post` privilege from all non-administrative users. Apply ACLs: Configure REST API Access Control Lists (ACLs) and/or firewall rules to restrict access to API endpoints exclusively to trusted administrator and HA peer IP addresses.
Multiple security vulnerabilities have been identified in OpenEXR, an open-source high-dynamic-range image file format library used for high-quality image processing and storage. These vulnerabilities include memory corruption, denial of service, out-of-bounds write, and information disclosure issues triggered by processing specially crafted EXR image files. Red Hat has issued an important security advisory for OpenEXR in Red Hat Enterprise Linux 9, providing updated packages to address these issues. The vulnerabilities affect various architectures supported by Red Hat Enterprise Linux 9 and related products. No CVSS scores are provided, but the overall severity is rated as important by Red Hat. Users are advised to apply the available updates as detailed in the Red Hat advisory.
A threat actor used AI tools with disabled safety controls to plan and execute multiple cybercrime activities, including credential theft, VoIP fraud, Wi-Fi password cracking, cryptocurrency reconnaissance, and database exfiltration. The actor targeted several Indian organizations, including universities, successfully exfiltrating thousands of personal records. AI-generated scripts and reconnaissance reports facilitated attacks such as WordPress brute force and MongoDB database theft. The actor also developed AI-powered VoIP fraud infrastructure and maintained automated pipelines for other criminal activities. Multiple unrelated actors are adopting similar AI-enabled toolchains, indicating a growing trend in AI-assisted cybercrime.
Unrestricted Upload of File with Dangerous Type vulnerability in Kitae Park Mang Board WP mangboard allows Upload a Web Shell to a Web Server.This issue affects Mang Board WP: from n/a through 2.4.4.
Missing Authorization vulnerability in bjorntech BjornTech Swish for WooCommerce woo-swish-e-commerce allows Exploiting Incorrectly Configured Access Control Security Levels.This issue affects BjornTech Swish for WooCommerce: from n/a through 3.7.8.
Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in Dylan Kuhn Geo Mashup geo-mashup allows Reflected XSS.This issue affects Geo Mashup: from n/a through 1.13.22.
Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in Brainstorm Force SureCart surecart allows Stored XSS.This issue affects SureCart: from n/a through 4.9.5.
Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') vulnerability in Bold Grid weForms weforms allows Path Traversal.This issue affects weForms: from n/a through 1.6.28.
Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in Wasiliy Strecker Contest Gallery contest-gallery allows Stored XSS.This issue affects Contest Gallery: from n/a through 34.0.1.
Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in ThemeHunk ThemeHunk Customizer themehunk-customizer allows Reflected XSS.This issue affects ThemeHunk Customizer: from n/a through 2.8.7.
Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in VideoWhisper.com Paid Videochat Turnkey Site ppv-live-webcams allows Reflected XSS.This issue affects Paid Videochat Turnkey Site: from n/a through 7.5.14.
