Popular GitHub Repositories for Safeguards

This is an open-source version of the representation engineering framework for stopping harmful outputs or hallucinations on the level of activations. 100% free, self-hosted and open-source.

First-of-its-kind AI benchmark for evaluating the protection capabilities of large language model (LLM) guard systems (guardrails and safeguards)

Stdio MCP Server wrapping custom Python runtime (LocalPythonExecutor) from Hugging Faces' `smolagents` framework. The runtime combines the ease of setup (compared to docker, VM, cloud runtimes) while providing safeguards and limiting operations/imports that are allowed inside the runtime.

Field research exposing how LLM safeguards collapse under polite, persistent interaction. Includes full report, metrics, session logs, and the AION conditioning protocol.

This repository contains the data and Python script used to generate insights and a taxonomy of language model safeguards bypassing techniques within Generative AI (Gen-AI) systems, as presented in the research paper "Taxonomy of Gen-AI Jailbreaks."

AI Agent Data Isolation System A universal framework to prevent data contamination in AI agent environments using vector databases. This system safeguards data integrity by detecting and filtering AI-generated content, applying session-based isolation, and preserving original data through configurable contamination detection patterns.