Wasmtime versions 47.0.0 before 48.0.6 and 49.0.0 before 49.0.3 have a memory misalignment flaw in the VMContext layout. This flaw causes undefined behavior when WebAssembly modules contain an odd number of exception tags, misaligning subsequent host fields in embedders with exceptions enabled. No exploitation path is currently known, and the issue has a low severity rating.
Wasmtime versions 46.0.0 before 48.0.6 and 49.0.0 before 49.0.3 contain a vulnerability that causes a panic (crash) when the host reads a component model stream as a different Rust item type than it was originally created with. This occurs because the host-to-host read path asserts matching TypeIds instead of returning an error, allowing guest code to trigger a crash by returning a host-created StreamReader<U> that the embedder reads as StreamReader<T>.
A vulnerability in Wasmtime before versions 36.0.18, 48.0.6, and 49.0.3 allows unbounded host memory growth. The issue arises because resource handle tables are capped only per component instance and do not account for allocations charged to any ResourceLimiter or StoreLimits. This enables a malicious guest to repeatedly allocate resources and instantiate sub-components, bypassing intended limits.
ZITADEL versions 3.0.0 through 3.4.15 and 4.0.0 before 4.19.5 have a missing authorization vulnerability in the Login V2 component. This flaw allows attackers to approve OAuth 2.0 device authorizations using a victim's active session without their consent. By sending a signed-in user a specially crafted link containing the attacker's user code, the attacker can obtain the victim's tokens and bypass project access controls via the OIDC device authorization API.
ZITADEL versions from 4.0.0 up to but not including 4.19.5 contain a host header injection vulnerability in the Login V2 UI. This flaw allows unauthenticated attackers to manipulate certain HTTP headers to poison password-reset, invite, and email-verification links. By exploiting this, attackers who know a victim's login name can redirect these links to their own host, intercept single-use codes, and potentially take over user accounts, including those of instance administrators.
Parse Server versions before 8.6.106 and 9.0.0 before 9.10.5-alpha.7 improperly handle dot notation keys with null or undefined values during object saving. This causes writes to be stored but return errors, skipping afterSave triggers and LiveQuery events. An attacker with write access can exploit this behavior to bypass audit logging, notifications, and updates to derived data.
Crawl4AI versions before 0.9.5 have a denial of service vulnerability in the Docker API server. The vulnerability arises because the _clamp_untrusted() function does not properly limit certain delay parameters in untrusted CrawlerRunConfig inputs. This allows unauthenticated attackers to send excessively large values to specific endpoints, causing resource exhaustion and blocking other crawling operations.
Parse Server versions before 8.6.107 and 9.0.0 before 9.10.5-alpha.8 have a vulnerability where the stored value of a field reverted by a beforeSave trigger is returned in object update responses. This occurs even if the client lacks read access to that object or field. An authenticated client who can update an object with such a beforeSave trigger can read protected values, including raw MFA authData when protectedFieldsTriggerExempt is enabled.
ImageMagick versions before 7.1.2-34 have a heap-based buffer overflow vulnerability in the -morph image processing feature. This flaw allows crafted images to cause writes beyond the bounds of a heap buffer, potentially crashing the application when processed with the -morph option.
ImageMagick versions before 7.1.2-34 have a stack overflow vulnerability in the TIFF decoder that can be triggered when the DNG decoder is disabled. This vulnerability allows an attacker to cause a denial of service by supplying a crafted image file that crashes the application.
Crawl4AI versions from 0.9.0 up to but not including 0.9.5 contain a JavaScript injection vulnerability in the Docker API server. This flaw allows authenticated API clients to execute arbitrary JavaScript on the crawl page by exploiting the allowlisted css_selector and wait_for fields. Attackers can inject quotes into css_selector or use js: conditions in wait_for to bypass the js_code block, altering the page state and the crawl content returned.
Crawl4AI versions prior to 0.9.5 have a blind server-side request forgery (SSRF) vulnerability in the Docker server API. This flaw allows API clients to bypass the egress proxy by enabling the check_cache_freshness feature. Attackers can exploit this by caching a controlled URL and then triggering the CacheValidator to follow 302 redirects or perform DNS rebinding to internal or cloud metadata service addresses, facilitating port and service discovery.
Crawl4AI versions from 0.9.0 up to but not including 0.9.5 have a vulnerability where the CosineStrategy model_name field in Docker API request bodies is not properly filtered. This allows authenticated clients to specify which HuggingFace model the server loads, potentially referencing malicious repositories. Exploitation can lead to arbitrary Python code execution under PyTorch versions below 2.6 or resource exhaustion via disk and memory consumption.
An unauthenticated sensitive data exposure vulnerability exists in WP Crowdfunding versions up to and including 2.2.2. This flaw allows attackers to access sensitive information without authentication.
An unauthenticated Cross Site Scripting (XSS) vulnerability exists in GlobalPayments WooCommerce versions up to and including 1.23.3. This vulnerability allows attackers to inject malicious scripts without authentication, potentially impacting confidentiality, integrity, and availability of affected systems.
A privilege escalation vulnerability exists in Buddyboss Platform versions up to and including 3.5.0, allowing users with subscriber-level access to escalate their privileges. This vulnerability is identified as CVE-2026-118082 and has a high severity rating with a CVSS score of 8.8.
An Insecure Direct Object Reference (IDOR) vulnerability affects Hotel Booking versions up to and including 3.8. This flaw allows an authenticated user with limited privileges to access subscriber data directly without proper authorization checks. The vulnerability has a CVSS 3.1 base score of 6.5, indicating a medium severity impact primarily on confidentiality.
Crawl4AI versions before 0.9.5 have an improper certificate validation vulnerability due to Chromium launching with the --ignore-certificate-errors flag. This allows network attackers performing man-in-the-middle or DNS-based attacks to intercept crawler traffic, serve forged content, and access cookies and headers. Additionally, a hard-coded --no-sandbox flag disables sandbox protections.
A PHP Object Injection vulnerability exists in FV Flowplayer Video Player versions up to and including 8.3.4. This vulnerability allows a user with shop manager privileges to potentially execute arbitrary code or cause other impacts by injecting crafted PHP objects.
A PHP Object Injection vulnerability exists in 10Web Social Photo Feed versions up to and including 1.5.6. This vulnerability allows an attacker with low privileges to execute arbitrary code remotely without user interaction, potentially leading to full system compromise.