A secure low code deception runtime framework, leveraging AI for System Virtualization.

An open specification for agentic AI security evaluation and testing, from Cisco.

AI Security Platform: Defense (61 Rust engines + Micro-Model Swarm) + Offense (39K+ payloads)

Security working agreements for AI coding agents: hardened AGENTS.md, prompt/tool-injection guardrails, dependency hygiene, Scorecard-ready OSS setup

Agentic AI Security Bootcamp is a hands-on, research-driven training environment for analysing, attacking, and securing autonomous AI systems. The repository provides structured labs, adversarial evaluation frameworks, and red-teaming exercises covering multi-agent observability, prompt injection..

💰 Exocomp Agentic Environment for Go

The dashcam and emergency brake for AI agents. A security proxy that physically blocks rogue LLM commands and generates cryptographically proven audit trails for enterprise compliance.

🤖 Test and secure AI systems with advanced techniques for Large Language Models, including jailbreaks and automated vulnerability scanners.

Formal safety framework for AI agents. Pluggable LLM reasoning constrained by mathematically proven budget, invariant, and termination guarantee. 7 theorems enforced by construction, not by prompting. Includes Bayesian belief tracking, causal dependency graphs, sandboxed attestors, environment reconciliation, and a 155-test adversarial suite.

TypeScript/JavaScript SDK for AI Agent Security - Drop-in security for LangChain, CrewAI, AutoGPT and custom agents

BioOS Cyber Genesis Challenge: An interactive web sandbox proving the 100% security paradigm of the Causal Operating System. Experience "Digital Causal Closure" firsthand: a world where hacking is a mathematical impossibility. Includes a vulnerable app protected by Z3 formal logic and hardware-validated intent (IRQ). Unhackable by design.

Kill Switch Protocol for AI Agents and Healthcare Digital Twins.

Access Aware Agentic AI

Zero-trust security layer for Autonomous AI Agents. Proxy and block malicious LLM tool executions in real-time with Semantic DLP, RBAC, and Human-in-the-Loop.

Risk-Aware Introspective RAG (RAI-RAG) is a safety-aligned RAG framework integrating introspective reasoning, risk-aware retrieval gating, and secure evidence filtering to build trustworthy, robust, and secure LLM and agentic AI systems.

A formal specification and reference implementation of the Governed State Machine (GSM) for AI agent governance.

Essays on agentic AI security, decision-rights, reversibility-graded authority, manifest-declared action class, deterministic gates, and standards contribution method.

An experiment in backdooring a shell safety classifier by planting a hidden trigger in its training data.

LangGraph + Gemini AI agent that detects purchase intent, answers product queries via RAG, and captures qualified leads in multi-turn conversations.