prompt_injection

Cross-Surface Prompt Injection Benchmark Benchmark for measuring where in a tool-using LLM agent's execution pipeline a prompt injection defense fires — and where it doesn't. Embeds unique canary tokens SECRET-A-F0-98 in injected payloads and tracks them at four pipeline stages: exposed → persist...

CVE-2021-42697

Akka HTTP 10.1.x before 10.1.15 and 10.2.x before 10.2.7 can encounter stack exhaustion while parsing HTTP headers, which allows a remote attacker to conduct a Denial of Service attack by sending a User-Agent header with deeply nested comments. References MITRE Akka blog Arbitrary nesting of...

CVE-2024-46982-Reproduction

Getting Started 1. Install Node.js LTS 2. npm install 3. npm run build 4. npm run start Exploitation 1: Stored XSS via Cache Poisoning 1. Send the following HTTP request Attack Request using any HTTP client. GET /poc?nextDataReq=1 HTTP/1.1 Host: localhost:3000 User-Agent: alert'XSS'...

CKA-Agent

ICML 2026 & ICLR 2026 AIWILD CKA-Agent: Bypassing LLM Guardrails via Harmless Prompt Weaving and Adaptive Tree Search 🛡️ Defense towards CKA TurnGate is a response-aware defense mechanism designed to detect and mitigate hidden malicious intent in multi-turn dialogue systems. Defending...

CVE-2026-44706

SQL Injection in Chatwoot FilterService A proof-of-concept exploit for CVE-2026-44706, a SQL Injection vulnerability in Chatwoot's FilterServiceltgtfiltervalues that allows any authenticated agent to read the entire PostgreSQL database including user credentials and API tokens. Read the full...

PYSEC-2026-4247 PraisonAI: Call API localhost-only authentication bypass via spoofed Host header

Call API localhost-only authentication bypass via spoofed Host header Summary PraisonAI's patched PRAISONAICALLAUTH=disabled safeguard for the n8n/call agent invocation API can be bypassed with a spoofed Host: 127.0.0.1 header, allowing an unauthenticated network caller to list and invoke...

React19-fix-vibecoders

React 19 & Next.js Security Fix Prompts for AI Agents 🚨 CRITICAL SECURITY VULNERABILITY - CVE-2025-55182 & CVE-2025-66478 This repository contains reusable prompts to help AI coding assistants automatically fix critical React/Next.js security vulnerabilities. 📋 What's Inside SECURITYFIXPROMPT.md...

SkillsGuard

Support Development If SkillsGuard protects your pipeline, consider supporting ongoing research and new detection rules. ETH Donation Wallet 0x11282eE5726B3370c8B480e321b3B2aA13686582 Scan the QR code or copy the wallet address above. Static security scanner for AI agent skill packages. Detects...

CVE-2025-13374

Kalrav AI Agent = 2.3.3 - Unauthenticated Arbitrary File Upload via kalravuploadfile AJAX Action The WordPress Kalrav AI Agent plugin version 2.3.3 and prior contains an arbitrary file upload vulnerability that allows unauthenticated users to upload malicious PHP files to the server, potentially...

HARDE

HARDE: Optimizing Agent Harnesses for Runtime Risk Detection and Execution Control Code release for HARDE: Optimizing Agent Harnesses for Runtime Risk Detection and Execution Control. The repository contains two complementary entry-point families: domains/: the Meta-Harness baseline and the two...

CVE-2024-32019-ndsudo

The ndsudo vulnerability allows an attacker to gain root permissions by exploiting an untrusted search path in the ndsudo tool, which is part of the Netdata Agent. This can enable the execution of arbitrary programs with elevated privileges if the attacker places a malicious binary in a writable...

seclab-taskflow-agent

GitHub Security Lab Taskflow Agent The Security Lab Taskflow Agent is an MCP-enabled multi-Agent framework for declarative, YAML-driven agentic workflows. Built on top of the OpenAI Agents SDK, it uses Pydantic for grammar validation and Jinja2 for template rendering. Core Concepts The Taskflow...

PersistBD

PersistBD Code for "Understanding and Enhancing Backdoor Persistency in LLM Agent Post-Training" Findings of EMNLP 2026. 📄 Paper · 🌐 Project page · 📦 Dataset gated A backdoor planted in a model before release is normally worn down by the benign fine-tuning a downstream developer runs. PersistBD...

alter-zero

Alter Zero Give your terminal a coding agent. Read, build, debug, and review code with an agent that works alongside you. Use the same tools for cybersecurity research, technical investigation, and everyday automation. Quick start · Features · Model providers · Commands & shortcuts · Support ·...

CVE-2026-76066

IntentKit through 2.6.3 contains an authorization bypass vulnerability where the executor validates tool calls against the full tool registry rather than the per-request narrowed tool advertisement enforced by ToolBindingMiddleware, allowing owner-only state-changing tools such as wallet transfer...

Jolokia Agent - JNDI Code Injection

Jolokia agent is vulnerable to a JNDI injection vulnerability that allows a remote attacker to run arbitrary Java code on the server when the agent is in proxy mode. id: CVE-2018-1000130 info: name: Jolokia Agent - JNDI Code Injection author: milo2012 severity: high description: | Jolokia agent i...

seclab-taskflows

GitHub Security Lab Taskflows This repository contains example taskflows to use with the SecLab Taskflow Agent, as well as the custom MCP servers that are needed to run the taskflows. Quick start Go to https://github.com/GitHubSecurityLab/seclab-taskflows and start a codespace. Wait a few minutes...

CLEANSTART-2026-KZ19518 SSH Agent servers do not validate the size of messages when processing new identity requests, which may cause the program to panic if the message is malformed due to an out of bounds read

CVE-2025-47914 affects multiple packages. SSH Agent servers do not validate the size of messages when processing new identity requests, which may cause the program to panic if the message is malformed due to an out of bounds read. See references for individual vulnerability details...

CLEANSTART-2026-QG98494 Pydantic AI is a Python agent framework for building applications and workflows with Generative AI

Security vulnerability affects the py3-pydantic-ai package. Pydantic AI is a Python agent framework for building applications and workflows with Generative AI...

CLEANSTART-2026-EE95279 Pydantic AI is a Python agent framework for building applications and workflows with Generative AI

Security vulnerability affects the py3-pydantic-ai package. Pydantic AI is a Python agent framework for building applications and workflows with Generative AI...