devpi-server versions before 6.21.0 have a path traversal vulnerability that allows authenticated users to write files outside the designated storage directory. This is achieved by creating a user and index named '..', enabling attackers to upload files with arbitrary names and contents outside the intended server directory.
Cesanta Mongoose versions 7.21 up to but not including 7.24 contain an out-of-bounds read vulnerability in the rx_ip6() function during IPv6 checksum validation. This flaw allows remote unauthenticated attackers to read memory beyond the bounds of received packets by sending specially crafted IPv6 packets with extension headers. The vulnerability arises because tcp6csum_ok() and icmp6csum_ok() functions use the full payload length for checksum validation, leading to reading adjacent receive-buffer memory.
Cesanta Mongoose versions 7.22 up to but not including 7.24 with the built-in TLS stack have an out-of-bounds read vulnerability in the mg_timegm() function. This flaw allows remote attackers to cause a denial of service by crashing the process during TLS handshake when presented with a crafted X.509 certificate containing invalid month values in the certificate validity period.
A path traversal vulnerability exists in hankcs HanLP up to version 2.1.6 in the uncompress function of the TAR Archive Extraction Handler component. This flaw allows remote attackers to manipulate file paths during archive extraction, potentially leading to unauthorized file access or modification. Exploit code has been published. Upgrading to version 2.1.7 resolves this issue.
probe-image-size version 7.4.0 has a regular expression denial of service (ReDoS) vulnerability in the lib/common/svg_header.js file. This flaw allows remote unauthenticated attackers to cause event loop blocking by supplying specially crafted SVG files with an unclosed internal subset followed by many ']' characters. The vulnerability triggers quadratic backtracking in probe.sync, probe(stream), or probe(url), leading to repeated stalls in the event loop.
probe-image-size versions before 7.5.0 have an algorithmic-complexity denial of service vulnerability in the lib/exif_utils.js component. This flaw allows remote unauthenticated attackers to exhaust CPU resources by submitting specially crafted JPEG images with self-referencing EXIF IFD pointers, causing the Node.js event loop to block for approximately 23 seconds.
OceanWP WordPress Theme versions before 4.2.9 contain a stored cross-site scripting (XSS) vulnerability in the oceanwp_edd_terms_list() function. This flaw allows authenticated users with Easy Digital Downloads (EDD) download category creation permissions to inject malicious JavaScript by crafting a specially formed category name. The injected script executes in the browsers of visitors who interact with the affected category link on the public EDD archive page, potentially compromising administrator sessions if they view the page while authenticated.
A vulnerability in highlight.js up to version 11.12.0 affects the hljs.highlight function in the livescript language grammar handler. Manipulating the text argument can cause inefficient regular expression complexity, potentially leading to denial of service. The flaw can be exploited remotely and public exploit code is available. A fix has been proposed but not yet merged.
A deserialization vulnerability exists in hankcs HanLP versions up to 2.1.6 in the function RNNLanguageModel.load_language_model. This flaw allows remote attackers to manipulate the model_file argument, potentially leading to unauthorized code execution or other impacts. The vulnerability is fixed in version 2.1.7.
HCL BigFix MCM has an improper implementation of the Content Security Policy (CSP) nonce mechanism. This flaw allows a remote attacker to weaken the CSP protection by exploiting an orphaned nonce value in the style-src directive of the CSP header, which is not matched by any nonce attribute on <style> or <script> elements. The vulnerability has a low severity score and does not impact confidentiality, integrity, or availability significantly.
The npm package 'aia-b-locator-fe' version 1.0.3 has been identified as malicious by the OpenSSF Package Analysis project. The package is flagged because it communicates with a domain associated with malicious activity. No further technical details or exploit information are provided.
Budibase versions up to 3.38.1 contain a vulnerability in self-hosted single-tenant deployments where the public bootstrap endpoint POST /api/global/users/init can be exploited after initial provisioning. An attacker can bypass the one-time initialization guard by supplying an arbitrary tenantId, causing the system to create new CouchDB databases and global admin users without authentication or CSRF protection. This allows unauthenticated creation of persistent admin accounts and potentially unlimited databases, leading to resource exhaustion and denial-of-service against the CouchDB datastore.
Budibase versions up to 3.38.1 contain a stored cross-site scripting (XSS) vulnerability that allows a low-privileged user to inject malicious HTML/JavaScript into application data. When a builder or admin later views this data in the Budibase Builder interface, the payload executes in their high-privilege session due to broken HTML escaping and unsafe rendering of user input. The Builder interface lacks a Content-Security-Policy header, enabling unrestricted script execution. This vulnerability enables full workspace takeover by escalating from a low-privileged user to builder/admin privileges.
The wcmatch Python library versions prior to 11.2 contain a Regular Expression Denial of Service (ReDoS) vulnerability when matching untrusted file names against patterns with multiple '*' wildcards using default flags. This causes polynomial-time complexity in rejecting crafted file names, leading to excessive CPU usage and potential blocking for minutes on a single core. The issue arises because each '*' wildcard is translated into a lazy match that allows the regex engine to try many splits, unlike the standard library's fnmatch which uses atomic groups to prevent this. The vulnerability affects scenarios such as upload handlers and directory scans where untrusted file names are matched against multi-star patterns. A fix is available in version 11.2 that changes the regex to use atomic groups for all but the last star segment, drastically improving performance and preventing the ReDoS.
Inscriptis versions from 2.0.0 up to but not including 2.7.6 have a vulnerability causing quadratic memory growth when processing multiple sibling <table> elements in the get_text() function. This leads to excessive memory consumption and large output sizes, potentially exhausting system memory with crafted HTML input. The issue arises from improper handling of indentation prefixes in block-level parent elements. A patch is available in version 2.7.6 that fixes this by correctly removing extra padding entries, significantly reducing memory usage and output size.
The svglib library's svg2rlg() function expands SVG <use> references recursively without limiting the total expansions, except for direct cycles. This allows crafted SVGs with nested <use> elements to cause exponential growth in rendering work, leading to high CPU and memory usage. This vulnerability can cause denial of service when rendering untrusted SVG content, including via xhtml2pdf which depends on svglib. A fix is available in svglib 2.3.1 that limits the number of <use> expansions to prevent resource exhaustion.
CVE-2026-104884 is a stored cross-site scripting (XSS) vulnerability in the Contao core-bundle affecting the frontend search results when using the legacy HTML5 search template. The vulnerability arises because the legacy template outputs a raw, unescaped link variable that can contain malicious script injected by a backend page editor. This script executes in the browsers of anonymous visitors and administrators viewing the search results. The default Twig-based search template is not affected as it properly escapes output. The issue is fixed in Contao versions 5.3.50 and 5.7.12.
A vulnerability in contao/core-bundle allows delegated non-admin user managers to escalate privileges by self-granting the 'theme_import' permission, which enables arbitrary PHP code execution on the server. This bypasses protections introduced in version 5.7.3 that restricted non-admin users from accessing the 'tpl_editor' module. The flaw affects versions prior to 5.3.50 and between 5.4.0-RC1 and 5.7.12, and it has been fixed in versions 5.3.50 and 5.7.12.
A vulnerability in contao/core-bundle's CSV import wizard allows non-admin editors with limited page/form scope to overwrite content records outside their authorized scope. This improper access control flaw enables integrity violations by bypassing per-record access checks, potentially leading to blind content overwrites. The issue affects versions from 5.0.0 up to but not including 5.3.50 and from 5.4.0-RC1 up to but not including 5.7.12. Fixed versions are 5.3.50 and 5.7.12.
A vulnerability in contao/newsletter-bundle allows a channel-scoped newsletter editor to bypass channel-scoped authorization. This enables unauthorized injection, mass mailing, and in some versions, deletion or denylisting of recipients in channels they should not have access to. The flaw arises because custom newsletter actions bypass the usual per-record permission checks, relying only on module-level access control.