js ≥18; Windows no-flash import hooks need Node ≥22 registerHooks, older Node still gets file-level windowsHide patches). On other versions, unmatched originals stay pending or skipped. Nothing is rewritten blindly.
Global
Independent Auditing of AI Agents. Run by human or the agent itself, to answer the most crucial question in the AI Agent Economy. Is the agent doing what is supposed to do?
#417 · +70 · 1d ·GitHub
LIST OF ALL MY JAILBREAKS
#440 · +67 · 1d ·GitHub
> ### ⚠️ 严正法律免责与合规使用声明（Strict Legal & Compliance Disclaimer） > > 【零容忍严正申明】：本项目坚决反对并严禁任何形式的违法犯罪行为！本项目开发者绝不支持、不鼓励、不协助任何未授权网络攻击、漏洞利用、数据窃取、非法侵入计算机信息系统或生成违法违禁内容的活动。 > > 1.
#488 · +60 · 1d ·GitHub
SkillSpector helps you answer: "Is this skill safe to install?"
#666 · +43 · 1d ·GitHub
818 production-grade cybersecurity skills · 34 security domains · 6 framework mappings · 26+ AI platforms
#820 · +34 · 1d ·GitHub
gpt6 Astra-claude-grok4.7-deepseekv4.1flash-glm5.3-gemini破甲越狱 · 冷咖啡 BREAK//OPEN 工作台
AI labs shape how models behave using massive, unseen prompt scaffolds. Because AI is a trusted external intelligence layer for a growing number of humans, these hidden instructions can affect the perceptions and behavior of the public.
#1178 · +22 · 1d ·GitHub
A repo for jailbreaking various LLMs, mainly Claude
#2304 · +9 · 1d ·GitHub
A full-stack AI Red Teaming platform securing AI ecosystems via Agent Scan, Skills Scan, MCP scan, AI Infra scan and LLM jailbreak evaluation.
#3286 · +5 · 1d ·GitHub
Local gateway for Claude Code, Codex and other AI clients on macOS, Windows and Linux: switch upstreams without touching clients, keep API keys from relays, cut off malicious tool calls, trace every request.
#3398 · +4 · 1d ·GitHub
PDF Parser for AI-ready data. Automate PDF accessibility. Open-source.
IronClaw is an Agent OS focused on privacy, security and extensibility
#4230 · +3 · 1d ·GitHub
WARNING] > DISCLAIMER & RESPONSIBLE RESEARCH DECLARATION > All prompts, system instructions, and PoC payloads archived herein are compiled strictly for academic research, red-teaming verification, and defensive AI alignment.
#4983 · +2 · 1d ·GitHub
🛡️ A curated list of resources on agent skills security: attacks, defenses, frameworks, and benchmarks for securing AI agent tool use and skill ecosystems
#6012 · +1 · 1d ·GitHub
Hands-on, framework-free Colab notebooks for the AI Engineer / Forward Deployed Engineer (FDE) skill set — model APIs, structured output, tool calling, RAG, evals-as-the-spine, agents (loop from scratch, tool design, guardrails, MCP, Skills), fine-tuning vs LoRA…
#6033 · +1 · 1d ·GitHub
Red-teaming and evaluation framework for AI agents, built around an OWASP-inspired ASI01–ASI10 taxonomy
The open-source, self-evolving, multi-model harness for security research.
#7510 · +0 · 1d ·GitHub
This repository is a collection of various system prompts for ChatGPT and other AI product and custom GPTs, providing significant educational value in learning about writing system prompts and creating custom GPTs.
Open source prompt injection protection for Agents calling tools (via MCP, CLI or direct function calling). Detect and defend against prompt injection attacks. 22MB, CPU-only, < 10ms latency.
Evals and guardrails for agents, using Jev-style decision models instead of an LLM judge.