Open Source Security Projects
Security is the practice of protecting systems, networks, applications, and information from unauthorized access, disruption, and misuse. It includes preventing attacks, identifying weaknesses, monitoring activity, and responding to incidents. Security tools help individuals and organizations reduce risk, safeguard privacy, control access, and maintain reliable services as threats and technology change.
Open source security tools range from vulnerability scanners and threat intelligence resources to encryption, network protection, access management, and isolated testing environments. When choosing a tool, consider its maturity, license, maintenance activity, documentation, system requirements, and compatibility with your existing workflows. Review how it handles sensitive data and whether it supports your threat model. These tools are useful to developers, security researchers, administrators, and anyone responsible for protecting digital systems.
64 repositories · updated October 3, 2026
Maskit: Local Privacy Gateway for LLMs and AI Tools
Maskit is a local privacy desensitization gateway engineered for large language models and AI tools. It automatically masks sensitive data in requests sent to AI services and then seamlessly restores it in streaming responses, ensuring private information remains local. This innovative solution supports various AI assistants like Cursor and Claude Code, along with any tool offering a configurable Base URL.
APort Agent Guardrails: Deterministic Pre-Action Authorization for AI Agents
APort Agent Guardrails provides deterministic pre-action authorization for AI agents, running security checks before any tool execution. This crucial mechanism prevents prompt injection from bypassing policy, ensuring robust and auditable protection for AI-driven operations.
Awesome AI Coding Sandboxes: A Security-First Guide to Agent Code Execution
Awesome AI Coding Sandboxes is a curated list of solutions for securely running code generated by autonomous AI agents. It uniquely ranks these sandboxes based on their security posture, focusing on isolation, egress control, and secrets handling. This resource helps developers choose robust environments to mitigate risks associated with untrusted agent code.
aport-spec: The Open Agent Passport (OAP) Specification for AI Agent Trust
The aport-spec repository introduces the Open Agent Passport (OAP) specification, a critical framework for establishing trust in AI agents. It defines a lightweight, cryptographically verifiable credential, enabling real-time, pre-action authorization for AI agents across various platforms. OAP provides the essential runtime trust layer for secure and scalable agentic commerce.
OpenBot: Secure AI Coworkers with Autonomous Browsers and Tools
OpenBot, an open-source project by CopilotKit, introduces AI coworkers that each operate within their own virtual computer, complete with a browser, files, and tools. This platform emphasizes secure and auditable execution, ensuring every action is decided and recorded before it happens. It offers a robust environment for autonomous operations, supporting any AG-UI agent framework.
ECC: The Agent Harness Performance Optimization System for LLMs
ECC is an advanced agent harness performance optimization system, providing a coordinated engineering system and toolbox for LLM development. It enables agents to plan, test, implement, review, and learn effectively, enhancing productivity and security for AI-driven coding workflows across platforms like Claude Code, Codex, and Cursor.
Agent Governance Toolkit: Policy Enforcement for Autonomous AI Agents
The Microsoft Agent Governance Toolkit (AGT) provides robust policy enforcement, zero-trust identity, and execution sandboxing for autonomous AI agents. It addresses critical security and compliance challenges, ensuring agents operate within defined boundaries and providing tamper-evident audit trails. AGT covers all 10 items of the OWASP Agentic Top 10, making it essential for shipping AI agents to production securely.
Agent Sandbox: Secure Local Development for AI Coding Agents
Agent Sandbox provides a robust and secure local development environment specifically designed for collaborating with AI coding agents. It ensures minimal filesystem access, configurable network egress policies, and secure secret injection, protecting your local machine from potentially risky agent operations. This project supports various AI agents and integrates seamlessly with both CLI and popular IDE devcontainer setups.
gh-aw-firewall: Secure Your Agentic Workflows with a Network Firewall
gh-aw-firewall is a robust network firewall designed specifically for agentic workflows, restricting outbound HTTP/HTTPS traffic to an allowlist of domains. It operates by running commands within a Docker sandbox, leveraging a Squid proxy for traffic filtering and an optional API proxy sidecar to securely manage LLM API keys. This project is a crucial component of GitHub's ongoing exploration into Agentic Workflows, enhancing their security and control.
OpenSandbox: A Secure and Extensible Sandbox Runtime for AI Agents
OpenSandbox is a powerful, general-purpose sandbox platform designed for AI applications. It provides secure, fast, and extensible runtime environments, supporting multi-language SDKs and Docker/Kubernetes deployments. This project is ideal for developing and evaluating AI agents in isolated, controlled settings.
CubeSandbox: Instant, Concurrent, and Secure Sandbox for AI Agents
CubeSandbox, developed by TencentCloud, is a high-performance, secure sandbox service built on RustVMM and KVM, designed specifically for AI agents. It offers ultra-fast startup times, hardware-level isolation, and high-density deployment, making it ideal for scalable and secure agent execution environments. The service is also fully compatible with the E2B SDK for seamless integration.
Destructive Command Guard (dcg): Protecting AI Agents from Dangerous Commands
The Destructive Command Guard (dcg) is a high-performance hook designed to prevent AI coding agents from executing dangerous git and shell commands. Built in Rust, dcg intercepts and blocks catastrophic operations like `git reset --hard` or `rm -rf`, offering clear explanations and safer alternatives. It provides a crucial safety layer for developers working with tools like Claude Code, Codex CLI, and GitHub Copilot.