Search This Blog

Powered by Blogger.

Blog Archive

Labels

Footer About

Footer About

Labels

Showing posts with label JSON. Show all posts

Fastjson Zero‑Day RCE Actively Targeting U.S. Companies

 

Hackers are abusing a critical Fastjson zero‑day remote code execution (RCE) flaw to compromise U.S. organizations by simply sending malicious JSON data to vulnerable Java applications. The attacks center on a vulnerability tracked as CVE‑2026‑16723 in Fastjson 1.x, an open‑source JSON parsing library widely used in Java and Spring Boot applications. 

The flaw affects versions 1.2.68 through 1.2.83 and is exploitable under default configurations, meaning attackers do not need user interaction, elevated privileges, or special deserialization gadgets to gain code execution on target servers. Threat intelligence firm ThreatBook first observed the malicious activity, with Imperva later confirming broad exploitation against firms in financial services, healthcare, computing, retail, and other industries. 

Technically, the issue lies in Fastjson’s type‑resolution logic, which performs attacker‑controlled resource lookups before enforcing AutoType restrictions. Security researchers from FearsOff showed that by abusing @type processing in Spring Boot “fat‑JAR” deployments, they could load and execute attacker‑supplied classes even when AutoType is disabled and without relying on traditional gadget chains. Alibaba’s advisory stresses that the vulnerability is exploitable on the most common Spring Boot executable JAR model launched via java -jar, and that specifying concrete classes for deserialization does not reliably mitigate the bug because payloads can hide inside generic Object or Map fields. 

What makes this zero‑day particularly dangerous is the combination of wide adoption, ease of exploitation, and lack of a direct patch for Fastjson 1.x. Imperva notes that most observed attacks currently hit U.S. organizations, with early signs in Singapore and Canada, and expects the campaign to expand globally as exploit code circulates. Fastjson 1.x is no longer actively maintained, and Alibaba has explicitly signaled that this branch is unlikely to receive a security fix, leaving thousands of applications exposed unless they are re‑engineered. 

Security vendors recommend that organizations immediately inventory applications using Fastjson 1.x and enable SafeMode via the appropriate JVM flag to hard‑block dangerous polymorphic deserialization paths. Longer term, developers are urged to either migrate to Fastjson2—which uses an allowlist‑first model and does not rely on the vulnerable @JSONType semantics—or replace Fastjson with other actively maintained JSON libraries. Until such changes are rolled out, defenders should deploy strict web application firewall rules to filter suspicious JSON payloads, monitor logs for exploitation attempts, and treat exposed services as high‑risk assets that may already be compromised.

Duolingo Data Breach: Hackers Posts Scrapped Data on Hacking Forum


After Discord’s data breach that resulted in its temporary halt in operations, the popular language learning app – Duolingo is facing a data breach.

An X post (previously tweeted) by user @vx-underground stated that a threat actor scraped data of over 2.6 million Duolingo users and posted it on the latest version of the hacking forum ‘Breached.’ BleepingComputer confirmed the breach in its recent post.

Apparently, the hackers gathered the data by manipulating existing vulnerabilities present in the Duolingo API, enabling access to user’s personal data, contact details, addresses, and much more, all by sending a valid email to the API.

The hackers further succeeded in finding active Duolingo users by feeding millions of email addresses to the vulnerable API. The email IDs were then used to create a dataset that contained public and non-public information. As an alternative, it is also feasible to supply a username to the API in order to obtain JSON output that contains sensitive user information.

But this is not the first time that this information has surfaced online. Falcon Feeds raised awareness of this problem via an X post in January. The scraped database was offered for sale for $1,500 on a previous iteration of the Breached hacker forum. Personal information about individuals, including email addresses, phone numbers, photographs, privacy settings, and much more, was revealed in the data.

Earlier, Duolingo had confirmed the data breach to TheRecord, assuring that it was investigating the issue. However, they did not mention that among the data was the private information of its users.

The most worrying aspect of this problem is that the corrupted API is still publicly accessible on the internet even though Duolingo first became aware of it in January. And, regrettably, this is not unexpected. Since most scraped data involves already-available information and is not the simplest to assemble into a credible threat, businesses frequently tend to ignore it.

In case of Duolingo, the breached data also involved sensitive data, that was not available publicly. While Duolingo is yet to address the issue, the most a user can do in this situation is modify their login credentials and/or delete their Duolingo accounts.