
Agentic-Evidence
OWASP reference project for agentic AI security evidence: correlates agent execution, identity, tool calls, and authorization decisions into…

OWASP reference project for agentic AI security evidence: correlates agent execution, identity, tool calls, and authorization decisions into…

Is 4-bit quantization more than noise to a chat model's safety behaviour? Code, pre-registration and labels.

[🎖️ Honorable Mention in HAIPS @COLM'26] The first benchmark to evaluate how well do LLM agents act in alignment with their assigned authorization…

Research code for poisoning attacks on the PGM-index, demonstrating how to craft adversarial data to degrade learned index performance.

Research pipeline that constructs verified successful experiences and organizes them into formation records to progressively poison model skills via…

Replication package studying how LLM vulnerability auditors defer to static-analysis evidence, with a tree-sitter C/C++ front end, controlled…

Jailbreak vulnerability across early-exit depths LLMs. Published at NeurIPS 2026 in the FLMSec workshop

End-to-end reproduction and cross-layer detection of CVE-2026-53576, the unauthenticated RCE in Kestra - taken past the base PoC to show how a common…

Example repository demonstrating how to exploit CVE-2026-77078 in multer, with usage instructions in English and Russian.

Proof-of-concept reproducer for Apache Camel camel-atmosphere-websocket dispatch header injection (CVE-2026-71300), demonstrating how an injected…

Fix-Like Artifacts With Embedded Defects

Detector + PoC for Linux page-cache write vulnerabilities: Copy Fail (CVE-2026-31431) and Dirty Frag (CVE-2026-43284/43500). Authorized security…

I Found a Zero-Day Vulnerability in langchain — Here’s How It Went

A curated list of public TEE resources for learning how to reverse-engineer and achieve trusted code execution on ARM devices

CVE-2025-54100 (CVSS 7.8 High) is a command injection vulnerability in the Invoke-WebRequest cmdlet of Windows PowerShell 5.1. It arises from…

This repository documents research into deserialization behavior within Next.js React Server Components (RSC) using the Flight protocol. It focuses…

CAWODOG is a proof-of-concept project demonstrating how to protect Python-based AI models deployed on offline industrial machines. Across three…

Artefacts for blog post on finding CVE-2025-37899 with o3