Offical repository for NeurIPS 2025 paper "From Judgment to Interference: Early Stopping LLM Harmful Outputs via Streaming Content Monitoring".

An interface library for RL post training with environments.

Claude Code for Financial Market

CLI security scanner built for the agentic era. Detects CI/CD misconfigs, agent permission risks, MCP tool injection, hardcoded secrets, and DMCA-flagged AI dependencies.

“ToolHazard: Scalable Red Teaming of LLM Agents via Adversarial Tool-Interactive Environment and Task Synthesis”

AI Security Scanner - Test your AI systems for prompt injection and extraction vulnerabilities

A debugging framework for agentic AI systems: diagnose failures, attribute root causes, recover with evidence, and validate fixes through reruns.

SafeLine is a self-hosted WAF(Web Application Firewall) / reverse proxy to protect your web apps from attacks and exploits.

Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content

Echo Agent 是一个可自托管、长期运行、持续学习的 AI Agent，面向个人与团队的私有自动化场景。它可以部署在自有服务器上，统一连接模型、工具、记忆、权限与消息入口。内置四层认知记忆、遗忘曲线与矛盾检测机制，能够在跨会话任务中持续沉淀上下文，并保持长期记忆的质量。针对命令执行、文件操作等高风险行为，它提供基于 LLM 的审批与解释机制，为关键操作建立可审计、可追溯的安全边界。原生…

The official implementation of the work "TopicAttack: An Indirect Prompt Injection Attack via Topic Transition"

POC管理平台已集成公开POC仓库内容,可自行添加POC到平台实现集中管理,快速搜索,验证

Code for paper "ActBench: Self-Evolving Benchmark of Behavioral Safety in Cowork Agents"

FinanceHarness: Autonomous Financial Deep Research Framework

The Official Repository for Paper "HarmfulSkillBench: How Do Harmful Skills Weaponize Your Agents?"

Installable AI coding workflow for risk-based routing, scoped sub-agents, and evidence-backed completion.

A lightweight CodeStable distribution

Silero VAD: pre-trained enterprise-grade Voice Activity Detector