AVideo: IDOR in AI Plugin Allows Stealing Other Users' AI-Generated Metadata and Transcriptions

Summary The AI plugin's save.json.php endpoint loads AI response objects using an attacker-controlled $REQUEST'id' parameter without validating that the AI response belongs to the specified video. An authenticated user with AI permissions can reference any AI response ID — including those generat...

PT-2026-27948

Name of the Vulnerable Software and Affected Versions Widget Wrangler versions prior to 2.3.9 Description A code injection issue exists in Jonathan Daggerhart Widget Wrangler. The issue involves improper control of code generation. This allows for code injection. Recommendations Update Widget...

CVE-2026-33527 Parse Server: Session update endpoint allows overwriting server-generated session fields

Parse Server is an open source backend that can be deployed to any infrastructure that can run Node.js. Prior to versions 8.6.57 and 9.6.0-alpha.48, an authenticated user can overwrite server-generated session fields such as expiresAt and createdWith when updating their own session via the REST...

EUVD-2026-14974

Parse Server's Session Update endpoint allows overwriting server-generated session fields...

CVE-2026-33335 Vikunja Desktop allows arbitrary local application invocation via unvalidated shell.openExternal

Vikunja is an open-source self-hosted task management platform. Starting in version 0.21.0 and prior to version 2.2.0, the Vikunja Desktop Electron wrapper passes URLs from window.open calls directly to shell.openExternal without any validation or protocol allowlisting. An attacker who can place ...

HCL AION Information Disclosure Vulnerability (CNVD-2026-15145)

HCL AION is an AI lifecycle management platform from HCL India. HCL AION suffers from an information disclosure vulnerability that stems from the predictability of certain identifiers, which can be exploited by an attacker to cause the attacker to infer or guess system-generated values, triggerin...

CVE-2026-32742 Parse Server session creation endpoint allows overwriting server-generated session fields

Parse Server is an open source backend that can be deployed to any infrastructure that can run Node.js. Prior to 9.6.0-alpha.17 and 8.6.42, an authenticated user can overwrite server-generated session fields sessionToken, expiresAt, createdWith when creating a session object via POST...

GHSA-5V7G-9H8F-8PGG Parse Server session creation endpoint allows overwriting server-generated session fields

Impact An authenticated user can overwrite server-generated session fields sessionToken, expiresAt, createdWith when creating a session object via POST /classes/Session. This allows bypassing the server's session expiration policy by setting an arbitrary far-future expiration date. It also allows...

HCL AION 安全漏洞

HCL AION is an AI lifecycle management platform from HCL India. HCL AION suffers from an information disclosure vulnerability that stems from the predictability of certain identifiers, which can be exploited by an attacker to cause the attacker to infer or guess system-generated values, triggerin...

EUVD-2026-11338

PingPong is a platform for using large language models LLMs for teaching and learning. Prior to 7.27.2, an authenticated user may be able to retrieve or delete files outside the intended authorization scope. This issue could result in retrieval or deletion of private files, including user-uploade...

CVE-2026-32097

PingPong, a platform for teaching/learning with LLMs, has a vulnerability prior to 7.27.2 where an authenticated user could retrieve or delete files outside the intended authorization scope. The issue allows retrieval of private files and deletion of files (including user uploads and model output...

Flowise Allows Mass Assignment in `/api/v1/leads` Endpoint

Summary A Mass Assignment vulnerability in the /api/v1/leads endpoint allows any unauthenticated user to control internal entity fields id, createdDate, chatId by including them in the request body. The endpoint uses Object.assign to copy all properties from the request body to the Lead entity...

Pakistan-Linked APT36 Floods Indian Govt Networks With AI-Made ‘Vibeware’

Bitdefender research reveals Pakistani group APT36 is using AI-generated vibeware and trusted cloud services like Google Sheets to target Indian officials...

CVE-2026-27641

Flask-Reuploaded provides file uploads for Flask. A critical path traversal and extension bypass vulnerability in versions prior to 1.5.0 allows remote attackers to achieve arbitrary file write and remote code execution through Server-Side Template Injection SSTI. Flask-Reuploaded has been patche...

EUVD-2026-8616

Flask-Reuploaded vulnerable to Remote Code Execution via Server-Side Template Injection...

A week in security (February 16 – February 22)

Last week on Malwarebytes Labs: Age verification vendor Persona left frontend exposed, researchers say Facebook ads spread fake Windows 11 downloads that steal passwords and crypto wallets AI-generated passwords are a security risk Intimate products maker Tenga spilled customer data Meta patents ...

Can You Tell It'S AI? Human Perception of Synthetic Voices in Vishing Scenarios

Large Language Models and commercial speech synthesis systems now enable highly realistic AI-generated voice scams vishing, raising urgent concerns about deception at scale. Yet it remains unclear whether individuals can reliably distinguish AI-generated speech from human-recorded voices in...

AI-generated passwords are a security risk

Using Artificial Intelligence AI to generate your passwords is a bad idea. It's likely to give that password to a criminal who can then use it in a dictionary attack—which is when an attacker runs through a prepared list of likely passwords words, phrases, patterns with automated tools until one ...

Code Injection

org.apache.avro, avro-compiler is vulnerable to Code Injection. The vulnerability is due to improper validation of untrusted Avro schemas during specific record generation, where attacker-controlled schema content can be incorporated into generated Java source code without sufficient sanitization...

Discovering Universal Activation Directions for PII Leakage in Language Models

Modern language models exhibit rich internal structure, yet little is known about how privacy-sensitive behaviors, such as personally identifiable information PII leakage, are represented and modulated within their hidden states. We present UniLeak, a mechanistic-interpretability framework that...