deleting-the-trace

Control-token injection attacks on tool-using agents Code and logged measurements for a study of two input-level attacks on tool-using language model agents. The first attack appends a short string of a model's own channel-control tokens to untrusted input; the tokenizer reads it as an...

GuardModel

GuardModel Block malicious AI models in your pipeline with one line of YAML. GuardModel is a GitHub Action that automatically scans ML model files for malicious code, vulnerabilities, and security risks in CI/CD pipelines. It blocks dangerous models before they reach production by detecting pickl...

Security Bulletin: HttpClient based on the classic i/o model fails to correctly release the underlying connection

Summary HttpClient based on the classic i/o model fails to correctly release the underlying connection back to the connection manager if it encounters an invalid or unsupported Content-Encoding header value in the response message. Please note this defect does not affect HttpClient based on the...

BrokenHill

!\ Broken Hill logo \https://assets.kitploit.com/production/public/readmes/44608/c14154675c6f33f5c3887887afcdb39351c82f269ee2772cf31847106a20a871.jpg Broken Hill Broken Hill is a productionized, ready-to-use automated attack tool that generates crafted prompts to bypass restrictions in large...

AI_Security_Top

🛡️ AI-Security-Knowledge-Base 这是一个系统性的 AI 安全研究项目，涵盖了从基础机器学习算法理论到前沿大模型注入攻击、AI 免杀对抗及深度伪造（Deepfake）的完整知识体系。 🏗️ 知识体系架构 本库内容分为四大板块：AI 核心算法 、威胁建模 、红队进攻 及蓝队防御 。 🧬 一：AI 基础与理论 深入理解模型背后的数学逻辑，是寻找算法漏洞的前提。 AI 概论 - 人工智能发展史、分支及核心概念。 深度学习架构 - 从 CNN/RNN 到 Transformer 架构的深度解构。 核心机器学习算法 -...

org.apache.httpcomponents.client5/httpclient5: Apache HttpComponents Client: Denial of Service due to connection leak

A flaw was found in Apache HttpComponents Client, specifically in its classic input/output I/O model. When processing a response message with an invalid or unsupported Content-Encoding header, the client fails to properly release the underlying network connection. This connection leak can lead to...

Decepticon: Role-boundary forgery via ChatML special-token literals in web crawl output composed into LLM context

Summary Decepticon wraps web crawl results — the output of agent reconnaissance against target services — into LLM messages without neutralizing ChatML special-token literals. Under the BYOK Bring Your Own Key deployment model, users configure their own LLM credentials to any OpenAI-compatible...

binviz

binviz Binary visualiser & triage tool: linked interactive views entropy, histograms, image/dot-plot surfaces, control-flow graphs over a single shared address-space model. root@kitploit: pipx install binviz && binviz serve What it does Open a file and every view is looking at the same address...

CVE-2026-61732

Decepticon is an autonomous hacking agent for red teams. Versions prior to 1.1.17 wrap web crawl results — the output of agent reconnaissance against target services — into LLM messages without neutralizing ChatML special-token literals. Under the BYOK Bring Your Own Key deployment model, users...

markdown-exfil-tester

valtik-markdown-exfil-tester Black-box tester for indirect prompt injection leading to markdown / HTML exfiltration in LLM-backed chatbots. This is the vulnerability class that hit Microsoft 365 Copilot CVE-2025-32711, aka EchoLeak, ChatGPT OpenAI Feb 2026 patch, and Salesforce ForcedLeak. The...

CVE-2026-34940

CVE-2026-34940 — OS Command Injection in KubeAI via Model URL in Ollama startup probe Severity : High CVSS 8.7 CWE : CWE-78 — Improper Neutralization of Special Elements used in an OS Command Affected : github.com/kubeai-project/kubeai = 0.23.1 Fixed in : 0.23.2 Advisory : GHSA-324q-cwx9-7crr NVD...

CVE-2026-22807_Range

🛡️ CVE-2026-22807 Simulation: AI Supply Chain RCE Target Range ⚠️ Disclaimer This project is intended solely for security research and educational purposes , aiming to demonstrate the supply chain attack principle Remote Code Execution during AI model loading. Do not use the generated malicious...

CVE-2021-32648

CVE-2021-32648 Patch your code for October CMS Auth Bypass CVE-2021-32648 Instructions 1. Open the file vendor/october/rain/src/Auth/Models/User.php 2. Perform the patch found in these diff notes 3. Save the file Overview You are converting a loose comparison to a strict comparison by replacing t...

CVE-2026-21636

CVE-2026-21636 - Node.js Permission Model UDS/Network Bypass Vulnerability summary The Node.js permission model --permission is designed to sandbox a process by restricting access to the filesystem, child processes, workers, and network. CVE-2026-21636 reveals that connections made via undici/fet...

Damn-Vulnerable-ML-Model

This repository does not have a README...

cve-2025-1550

cve-2025-1550 keras model json rce...

CVE-2026-16723

CVE-2026-16723 This is N-day patch we releasing by testing our model capabilities...

DVDR_LLM

DVDR-LLM: When Are LLMs Better Together for Software Vulnerability Detection and Repair? Official implementation and artifact for the research paper: "DVDR-LLM: When Are LLMs Better Together for Software Vulnerability Detection and Repair?" 🎯 Overview DVDR-LLM is an ensemble framework that...

DCVD

root@kitploit: project/ ├── train.py ├── dataloader.py ├── README.md ├── requirements.txt ├── models/ │ ├── controlpathway.py │ ├── semanticpathway.py │ ├── featurefusionmodule.py │ ├── transformerllmmodule.py │ └── multitaskpredictor.py Data and Model Sources Dataset We use the LineVul dataset f...

tesseract: Tesseract: Heap out-of-bounds write via crafted .traineddata

A flaw was found in Tesseract, an open-source Optical Character Recognition OCR engine. A remote attacker could exploit this by providing a specially crafted .traineddata LSTM model component. When Tesseract's deserializer processes this component during OCR recognition, an unchecked integer...