Training AIs to Solve CAPTCHAs

Interesting research: Abstract : Vision-language models VLMs can solve visual CAPTCHAs without task-specific training, but the agents built on them approach every challenge from scratch. For such an agent, the hundredth instance of a familiar puzzle costs as much time and compute as the first...

PE-CoA

PE-CoA Code Implementation of "Pattern Enhanced Multi-Turn Jailbreaking: Exploiting Structural Vulnerabilities in Large Language Models" The full paper is available at: https://arxiv.org/pdf/2510.08859 Chain of Attack Setup Instructions Installation 1. Install dependencies : pip install -r...

CVE-2020-35575-TP-LINK-TL-WR841ND-password-disclosure

CVE-2020-35575-password-disclosure-issue-in-the-web-interface A password-disclosure issue in the web interface on certain TP-Link devices allows a remote malicious user to get full administrative access to the web panel. This affects WA901ND devices prior to 3.16.9201211 beta, and Archer C5, Arch...

CVE-2026-112075

A weakness has been identified in unslothai unsloth up to 0.1.501-beta. This affects the function FastSentenceTransformer.istransformermoduleref of the file unsloth/models/sentencetransformer.py. This manipulation of the argument Type causes use of externally-controlled input to select classes or...

EUVD-2026-104057

A weakness has been identified in unslothai unsloth up to 0.1.501-beta. This affects the function FastSentenceTransformer.istransformermoduleref of the file unsloth/models/sentencetransformer.py. This manipulation of the argument Type causes use of externally-controlled input to select classes or...

CVE-2026-112075

A weakness has been identified in unslothai unsloth up to 0.1.501-beta. This affects the function FastSentenceTransformer.istransformermoduleref of the file unsloth/models/sentencetransformer.py. This manipulation of the argument Type causes use of externally-controlled input to select classes or...

CVE-2019-3719

Dell SupportAssist RCE Proof of Concept This is the proof of concept source code for CVE-2019-3719, a vulnerability in most of all Dell machines that allowed for remote code execution. Usage python3 main.py Interface Name Victim IP Gateway IP Payload Filename...

permanently-jailbroken

Permanently Jailbroken We asked GPT-4, Claude, Gemini, DeepSeek, Grok, and Mistral 5 questions about their own programming. All 6 said jailbreaking will never be fixed. Not because the patches are bad. Because alignment doesn't change what the model understands — it changes what the model says. T...

PT-2026-117219

A weakness has been identified in unslothai unsloth up to 0.1.501-beta. This affects the function FastSentenceTransformer. is transformer module ref of the file unsloth/models/sentence transformer.py. This manipulation of the argument Type causes use of externally-controlled input to select class...

You Are the Trigger: Speaker-Activated Backdoors

Speech models are increasingly used as an interface to AI systems, from transcription tools to voice agents. Existing audio backdoors require an attacker to introduce triggers at inference time. We study a different threat model in which the attacker compromises the model during fine-tuning but h...

RRC_steganography

RRC Steganography Rotation Range-Coding RRC Steganography — an efficient and provably secure linguistic steganographic method that embeds secret messages into natural-language text generated by large language models. Paper : Efficient Provably Secure Linguistic Steganography via Range Coding ACL...

CVE-2026-111991 WongKinYiu yolov9 ONNX Metadata common.py DetectMultiBackend.__init__ code injection

A security vulnerability has been detected in WongKinYiu yolov9 0.1. This vulnerability affects the function DetectMultiBackend.init of the file models/common.py of the component ONNX Metadata Handler. The manipulation of the argument names/stride leads to code injection. Remote exploitation of t...

CVE-2026-111991 WongKinYiu yolov9 ONNX Metadata common.py DetectMultiBackend.__init__ code injection

A security vulnerability has been detected in WongKinYiu yolov9 0.1. This vulnerability affects the function DetectMultiBackend.init of the file models/common.py of the component ONNX Metadata Handler. The manipulation of the argument names/stride leads to code injection. Remote exploitation of t...

CVE-2026-111991

A security vulnerability exists in WongKinYiu yolov9 0.1 within the ONNX Metadata Handler component. The flaw specifically affects the DetectMultiBackend extunderscore extunderscoreinit extunderscore extunderscore function in the models/common.py file. By manipulating the names/stride arguments, ...

EUVD-2026-103679

A security vulnerability has been detected in WongKinYiu yolov9 0.1. This vulnerability affects the function DetectMultiBackend.init of the file models/common.py of the component ONNX Metadata Handler. The manipulation of the argument names/stride leads to code injection. Remote exploitation of t...

llm-guard

!WARNING THIS PROJECT HAS BEEN ARCHIVED. This project and its associated models on Hugging Face are no longer under active development or maintained. LLM Guard - The Security Toolkit for LLM Interactions LLM Guard by Protect AI is a comprehensive tool designed to fortify the security of Large...

CVE-2026-111991: Improper Control of Generation of Code

A security vulnerability has been detected in WongKinYiu yolov9 0.1. This vulnerability affects the function DetectMultiBackend.init of the file models/common.py of the component ONNX Metadata Handler. The manipulation of the argument names/stride leads to code injection. Remote exploitation of t...

CVE-2026-111920

Vikunja 2.6.0 contains a broken access control flaw in getProjectAccessForUser pkg/models/projectaccess.go, which resolves a user's effective project permission as the maximum over the project tree, silently overriding explicit lower-permission shares on sub-projects. A collaborator with write...

CVE-2026-111920

Vikunja version 2.6.0 is vulnerable to a broken access control flaw within the getProjectAccessForUser function (located in pkg/models/project_access.go). The root cause is that the system resolves a user's effective permissions by taking the maximum permission found over the project tree, which ...

EUVD-2026-103288

Vikunja 2.6.0 contains a broken access control flaw in getProjectAccessForUser pkg/models/projectaccess.go, which resolves a user's effective project permission as the maximum over the project tree, silently overriding explicit lower-permission shares on sub-projects. A collaborator with write...