Open Source Security Projects
Discover 166 open source Security repositories from GitHub, each with an analysis of what it does, key features, use cases and alternatives. Security projects here are most often combined with Self Hosted, Python and AI Agents. Last updated October 4, 2026.
166 repositories · updated October 4, 2026
agent-sandbox: Manage Isolated Stateful Workloads on Kubernetes
Agent Sandbox adds Kubernetes APIs and controllers for isolated, stateful workloads with stable identity and persistent storage. Use it to run long-lived AI agent environments, development sessions, or other singleton workloads that need managed lifecycle control.
jinja: Render Dynamic Templates in Python
Jinja is a fast, extensible template engine that turns templates and application data into documents. Python developers use it for HTML and other generated text, with features such as inheritance, autoescaping, async support, and sandboxing.
hashids-python: Obfuscate Numeric IDs as Short Hashes
hashids-python encodes one or more integers as short, URL-friendly strings and decodes them back. It suits Python applications that want to make exposed database IDs less obvious, but it is not a security mechanism.
destructive_command_guard: Block Dangerous Commands Before Execution
dcg is a Rust hook that checks commands from AI coding agents before they run and blocks destructive operations. It suits developers who want guardrails against accidental data or work loss, with configurable coverage for additional tools and services.
awesome-web-security: Find Web Security Learning Resources
A curated directory of web security articles, tools, and references covering vulnerability classes, testing techniques, and browser security. Useful for security learners and practitioners who need a starting point for research or authorized testing.
MacOS-MCP: Automate macOS with AI Agents
MacOS-MCP is a Python MCP server that lets AI agents interact with macOS apps and interfaces through accessibility controls and system tools. It suits developers building desktop workflows, with macOS permissions and careful security review required.
memoripy: Build Evidence-Based Memory for AI Agents
Memoripy is a local memory runtime for AI agents that keeps versioned, sourced records and explains why they appear in recall. It suits developers who need scoped, auditable memory rather than an opaque store of accumulated text.
JailbreakEval: Compare LLM Jailbreak Evaluators
JailbreakEval brings together automated methods for assessing whether language-model responses comply with jailbreak attempts. Researchers can compare evaluators across datasets, while developers can build and benchmark new evaluation methods.
EasyJailbreak: Build and Evaluate LLM Jailbreak Attacks
EasyJailbreak is a Python framework for assembling and testing jailbreak methods against language models. It suits researchers and developers who need reusable attack components and a structured way to evaluate model responses.
guardrails: Validate LLM Inputs and Outputs
Guardrails is a Python framework for checking LLM inputs and outputs against configurable validators and for generating structured data. It suits teams building AI applications that need validation and error handling around model responses.
MarkLLM: Implement and Evaluate LLM Watermarking
MarkLLM is a Python toolkit for implementing, visualizing, and evaluating text watermarking methods for large language models. It helps researchers compare watermark detection, robustness, and text-quality effects through shared APIs and evaluation pipelines.
PYAS: Layered Endpoint Security for Windows
PYAS is a Windows endpoint security application that combines local machine-learning and YARA scanning with real-time monitoring, optional cloud analysis, and kernel-level controls. It suits security researchers and Windows users evaluating layered protection, with driver and remediation features best tested in an isolated environment.