CVE Notify
19.7K subscribers
4 photos
300K links
Alert on the latest CVEs

Partner channel: @malwr
Download Telegram
🚨 CVE-2026-19582
In binutils 2.46.1 and prior versions, a victim who opens a crafted PE file using binutils could execute arbitrary code unknowningly via a stack buffer overflow out of bounds write.

🎖@cveNotify
🚨 CVE-2026-78619
Punk::Plugin::TOTP versions before 0.05 for Perl accept another account's recovery code at the two-factor challenge because totp_use_recovery compares user identifiers numerically.

The helper searches the recovery model for the submitted code's digest alone, across every user's rows, so the ownership test that follows is the only thing binding a code to the account it was issued to. That test compares the row's user_id with the challenged user's id through Perl's integer coercion, and an identifier with no leading digits coerces to zero, so any two of them compare equal. User models keyed on a username, an email address or a UUID hit that case, and a numeric key compares as intended.

The challenge route feeds a submitted value to the helper once TOTP verification fails, so an attacker who knows a victim's password and holds a recovery code of their own passes the victim's second factor.

🎖@cveNotify
🚨 CVE-2026-78655
Punk::Plugin::TOTP versions before 0.05 for Perl allow the second-factor attempt limit to be reset by replaying an earlier session cookie because the challenge route counts failures in the session.

The POST handler on challenge_path keeps the failure count as tries inside the totp_pending record in the session, raising it on each rejected code and deleting the pending record once it reaches attempts, five by default. Punk::Session carries the session in a signed cookie unless the application declares a store, and keeps no server-side record, so an earlier value of the same session stays valid until the expiry stamped inside it. A client that saves the cookie before its failed attempts and presents it again gets the pending record back with its counter, and the limit never fires. The replayed record is accepted while its own expiry, pending_ttl seconds from the challenge and 300 by default, has not passed.

Sessions declared with a store are not affected: the pending record and its counter then live server-side.

The attempt limit does not bound guessing of the second factor, which is left to the per-address rate limit the plugin registers on the same path, 30 requests per 60 seconds.

🎖@cveNotify
🚨 CVE-2026-15916
Missing Authorization vulnerability in Drupal Drupal core allows Forceful Browsing. This issue affects Drupal core versions: from 0.0.0 to 10.6.13, from 11.3.0 to 11.3.14, from 11.4.0 to 11.4.4, from 0.0.0 to 11.0.*, from 0.0.0 to 11.1.*, from 0.0.0 to 11.2.*.

🎖@cveNotify
🚨 CVE-2026-15917
Improper Neutralization of Input During Web Page Generation ("Cross-site Scripting") vulnerability in Drupal Drupal core allows Cross-Site Scripting (XSS). This issue affects Drupal core versions: from 11.3.0 to 11.3.14, from 11.4.0 to 11.4.4, from 0.0.0 to 11.2.*.

🎖@cveNotify
🚨 CVE-2026-44476
Doorkeeper is an OAuth 2 provider for Ruby on Rails. In version 1.9.0, an attacker who knows only a dynamically registered client's client_id, which is public information, can authenticate as that client at the token endpoint and obtain an access token without providing its client_secret. This occurs because the Dynamic Client Registration feature creates applications with confidential: false hard-coded, even though the registration response returns a client_secret and advertises support for the client_secret_basic and client_secret_post authentication methods; since Doorkeeper treats a blank or missing secret as valid for non-confidential (public) clients, the secret is never verified. Only projects that have explicitly enabled Dynamic Client Registration, which is disabled by default, are affected. This issue is fixed in version 1.10.0.

🎖@cveNotify
🚨 CVE-2026-54757
Compliance-trestle (Trestle) is a Python SDK and command-line tool for managing OSCAL compliance documents. In versions before 3.12.4 and versions 4.0.0 through 4.0.3, Trestle is vulnerable to server-side template injection that can lead to remote code execution. This occurs because the MDCleanInclude and MDSectionInclude Jinja2 tags re-parse untrusted Markdown content as template source code using a non-sandboxed jinja2.Environment. An attacker who controls content that Trestle renders, such as a crafted workspace Markdown file, a third-party SSP document, or a YAML lookup-table value, can inject a Jinja2 expression that traverses Python object internals to execute arbitrary operating system commands in the context of the Trestle process. This issue is fixed in versions 3.12.4 and 4.1.0.

🎖@cveNotify