adversarial-testing

Open-source adversarial testing engine, SDK, and CLI for AI agents. Runs locally or against the Humanbound Platform.

The Self-Evolving Agent Ecosystem — Trading agents that evolve through Darwinian selection and adversarial self-play

AI Robustness Evaluation System

Open-source test harness for AI agents. Stress-test production agents with adversarial multi-turn scenarios in CI

Toolkit for AI whitehats, internal red teams, llm bug bounty hunters and mlops - Adversarial testing for AI, LLMs, and Agents.

LLM Adversarial Security Arena — Jailbreak → Detect → Defend → Verify

Open-source framework for building and testing LLM-powered applications: IRIS (single-agent orchestration), AETHER (declarative multi-agent systems), and AEGIS (adversarial security testing). Developed at MSU Denver's Community-Centered Computing (C3) Lab.

Mechanism-grounded taxonomy of 40 LLM jailbreak patterns across 10 categories. 8,000-trial bootstrap evaluation for the June 2026 frontier (Claude Opus 4-8, GPT-5.5, Gemini 3.5, DeepSeek V4). Every citation direct-WebFetch verified; refuted claims documented.

MCP server that wraps the xAI Grok CLI. Lets Claude Code, Cursor, Cline, and any MCP host use Grok as a peer code reviewer, adversary, and second-opinion consultant.

Production-ready Claude Code decision intelligence marketplace with adversarial routing, evidence-driven analysis, specialized agents, benchmarks, and automated validation.

Opt-in Codex Skill for practical coding and externally anchored delivery with elastic agent teams, task DAGs, isolated candidates, and one canonical writer.

Your coding agent says it works. Gopnik tries to prove it doesn't — against the code, the delivered revision, and the checks themselves.

Red-team your AI agents from any coding IDE. Adversarial security testing skills for Claude Code, Cursor, Codex, and 40+ agents.

Formal adversarial testing of LLM-generated industrial robot (URScript) code: ISO 10218-1:2025 safety + CWE security analysis, AST static watchdog, URSim runtime. Simulation-only.

A marketplace of Claude Code plugins for adversarial security and architectural code review.

Elenchus MCP Server - Adversarial verification system for code review

Multi-perspective code review council for Claude Code. 3 advisors by default, 10 agents in deep mode (Opus + Codex). Evidence chains, adversarial self-test, dual-path verdict. Based on Karpathy's LLM Council.

Benchmark LLM jailbreak resilience across providers with standardized tests, adversarial mode, rich analytics, and a clean Web UI.

Evolutionary adversarial testing framework for AI safety using quality-diversity search to discover interpretable, transferable vulnerabilities across LLMs. (ICLR 2026)

Official GitHub Actions for Humanbound — adversarial security testing for AI agents in CI.