Tag
#ssl
The malware, operated by China-backed cyberattackers, has been significantly fortified with new evasive and post-infection capabilities.
### Impact Nonce generation does not use sufficient entropy nor a cryptographically secure pseudorandom source (https://github.com/guzzle/oauth-subscriber/blob/0.8.0/src/Oauth1.php#L192). This can leave servers vulnerable to replay attacks when TLS is not used. ### Patches Upgrade to version 0.8.1 or higher. ### Workarounds No. ### References Issue is similar to https://nvd.nist.gov/vuln/detail/CVE-2025-22376.
This week on the Lock and Code podcast, we speak with Anna Brading and Mark Stockley about whether anywhere is safe from AI slop.
### Impact A vulnerability has been reported in Password Pusher where an attacker can copy the session cookie before a user logs out, potentially allowing session hijacking. Although the session token is replaced and invalidated upon logout, if an attacker manages to capture the session cookie before this process, they can use the token to gain unauthorized access to the user's session until the token expires or is manually cleared. This vulnerability hinges on the attacker's ability to access the session cookie during an active session, either through a man-in-the-middle attack, by exploiting another vulnerability like XSS, or via direct access to the victim's device. ### Patches Although there is no direct resolution to this vulnerability, it is recommended to always use the latest version of Password Pusher to best mitigate risk. ### Workarounds If self-hosting, ensure Password Pusher is hosted exclusively over SSL connections to encrypt traffic and prevent session cookies fr...
An issue was discovered in TCPDF before 6.8.0. If libcurl is used, CURLOPT_SSL_VERIFYHOST and CURLOPT_SSL_VERIFYPEER are set unsafely.
From Chinese cyberspies breaching US telecoms to ruthless ransomware gangs disrupting health care for millions of people, 2024 saw some of the worst hacks, breaches, and data leaks ever.
Relive the 90s web era! The Neuro Nostalgia Hackathon challenged teams to transform modern sites into retro masterpieces…
### Impact Unprivileged user accounts with at least one SSH key can read arbitrary files on the system. For instance, they could leak the configuration files that could contain database credentials (`[database] *`) and `[security] SECRET_KEY`. Attackers could also exfiltrate TLS certificates, other users' repositories, and the Gogs database when the SQLite driver is enabled. ### Patches Unintended Git options has been ignored for creating tags (https://github.com/gogs/gogs/pull/7872). Users should upgrade to 0.13.1 or the latest 0.14.0+dev. ### Workarounds No viable workaround available, please only grant access to trusted users to your Gogs instance on affected versions. ### References https://www.cve.org/CVERecord?id=CVE-2024-39933
An issue was identified in the `VmFd::create_device function`, leading to undefined behavior and miscompilations on rustc 1.82.0 and newer due to the function's violation of Rust's pointer safety rules. The function downcasted a mutable reference to its `struct kvm_create_device` argument to an immutable pointer, and then proceeded to pass this pointer to a mutating system call. Rustc 1.82.0 and newer elides subsequent reads of this structure's fields, meaning code will not see the value written by the kernel into the `fd` member. Instead, the code will observe the value that this field was initialized to prior to calling `VmFd::create_device` (usually, 0). The issue started in kvm-ioctls 0.1.0 and was fixed in 0.19.1 by correctly using a mutable pointer.
### Summary An **Improper URL Handling Vulnerability** allows an attacker to access sensitive local files on the server by exploiting the `file:///` protocol. This vulnerability is triggered via the **"real-browser"** request type, which takes a screenshot of the URL provided by the attacker. By supplying local file paths, such as `file:///etc/passwd`, an attacker can read sensitive data from the server. ### Details The vulnerability arises because the system does not properly validate or sanitize the user input for the URL field. Specifically: 1. The URL input (`<input data-v-5f5c86d7="" id="url" type="url" class="form-control" pattern="https?://.+" required="">`) allows users to input arbitrary file paths, including those using the `file:///` protocol, without server-side validation. 2. The server then uses the user-provided URL to make a request, passing it to a browser instance that performs the "real-browser" request, which takes a screenshot of the content at the given URL....