Why it matters
This framework empowers builders to create automated security workflows that operate across diverse tech stacks. Its open-source nature and flexibility in choosing LLMs allow for customized, private security solutions that can be integrated into CI/CD pipelines or driven by coding agents.

What changed

0security has been released as an open-source framework designed for building and deploying continuous security agents. Developed by the Swiss Applied AI & Cybersecurity Research Lab, this tool aims to automate the detection, proof, remediation, and monitoring of vulnerabilities across a broad spectrum of a tech stack. This includes source code, web applications, APIs, dependencies (npm, pnpm, Cargo, PyPI), binaries, cloud infrastructure, and even AI agents themselves. The framework supports multiple deployment methods, including a Web UI, CLI, CI integration, and an MCP (Multi-agent Communication Protocol) server, all designed to run on the user's own infrastructure. Installation can be done via a curl script for macOS and Linux, with experimental Windows support via an executable, and Docker images available. Source installations require Node.js 24+ and pnpm.

Key functionalities include:

  • Vulnerability Discovery: Scanning source code, web apps, APIs, MCP endpoints, and packages.
  • Dependency Management: 0 deps scan and 0 deps fix for various package managers.
  • Binary Analysis: Reverse engineering capabilities and kernel/memory-safety fuzzing.
  • Attack Surface Mapping: Identifying domains, JavaScript, cloud, and Entra ID posture.
  • Automated Remediation: Reproducing findings, generating patches, and verifying fixes.
  • Persistent Agents: Building 'Sentinels' with defined scopes, budgets, and memory for scheduled or event-driven security tasks.
  • AI Agent Testing: Conducting prompt-injection campaigns.
  • CI/CD Integration: Automating scans with JSON and SARIF output.

The framework emphasizes privacy by running on the user's infrastructure and only sending necessary context to the chosen LLM. It supports various LLM providers and local OpenAI-compatible servers, offering flexibility in model selection.

Why it matters for builders

For AI builders and cybersecurity professionals, 0security offers a robust, open-source platform to automate and enhance security testing and remediation processes. The ability to build custom security agents that can continuously monitor and fix vulnerabilities across code, infrastructure, and even other AI agents significantly reduces manual effort and potential human error. Its flexible architecture, allowing for the use of any LLM provider or local models, ensures that sensitive data remains within the builder's control, aligning with privacy and security best practices.

Practical impact

Builders can leverage 0security to integrate automated security checks directly into their development workflows. For instance, the CLI allows for quick scans of local repositories using commands like 0 scan source ./your-repo --runtime api --depth quick --cost-ceiling 2. For continuous integration, the 0 automation run --template security-research --target /absolute/path/to/repo --workspace /absolute/path/to/repo command can be used. The MCP server functionality, 0 mcp-server --automations --workspace /absolute/path/to/repo, enables more complex agent-to-agent interactions for security tasks. Developers can also explore building custom plugins and scan modules to extend the framework's capabilities, or contribute to the project's development by reporting bugs or improving documentation.

Caveats and source limits

The source material indicates experimental support for Windows installations. While the framework supports a wide range of functionalities, specific performance benchmarks or independent validation of its vulnerability detection and fixing capabilities are not detailed in the provided excerpts. The documentation for specific scan modules, supported models, and advanced MCP usage is referenced but not fully elaborated upon within the given text, suggesting that further exploration of the project's documentation and community resources would be necessary for a comprehensive understanding of its operational limits and advanced features.

Sources

Written with AI assistance from the linked sources; every claim below was checked against them automatically. How we produce articles.

Claim check: 8/8 supported claims - 8 evidence links - 100% avg confidence
  • 0security is an open-source framework for building security agents that find, prove, fix, and monitor vulnerabilities.supported - github.com
  • The agents can find vulnerabilities across code, binaries, dependencies, infrastructure, and AI agents.supported - github.com
  • 0security runs on user-provided infrastructure and supports chosen LLMs.supported - github.com
  • The framework includes a Web UI, CLI, CI integration, and MCP support.supported - github.com
  • Installation is available via a script for macOS/Linux, Docker, and source compilation.supported - github.com
  • 0security supports dependency scanning and fixing for npm, pnpm, Cargo, and PyPI.supported - github.com
  • The framework can perform prompt-injection campaigns against AI agents.supported - github.com
  • 0security is licensed under MIT or Apache-2.0.supported - github.com

Caveats

  • Single-source caution: verify critical details at the linked source.
Radar score 89/100 - how it was calculated
Reliability82
Freshness100
Novelty77
Technical89
Developer96
Ecosystem72
Confidence100
  • Reliability 82: GitHub metadata supports source trust
  • Freshness 100: Fresh GitHub release date
  • Novelty 77: Fresh GitHub release
  • Technical 89: Repository technical metadata
  • Developer 96: Developer tooling signals
  • Ecosystem 72: Fresh GitHub release
  • Confidence 100: Claims have reliable evidence
Share
XLinkedInHacker News

Discussion

Loading comments...

Related articles

Agents - Nov 15, 2026n8n Releases Nod: Open Source Personal Agentn8n has released Nod, an open-source personal agent designed to run locally on a user's machine. Nod can automate tasks like chasing invoices, preparing for calls, and canceling online trials by interacting with websites and applications.Agents - Nov 13, 2026agentes.chat: Open-Source Slack Alternative with AI Agentsagentes.chat is an open-source, self-hosted Slack alternative designed for integrated team and AI agent collaboration. It supports multiple LLMs, including Claude, ChatGPT, Gemini, and Grok, and offers features like agent-specific computers, browsers, routines, and approval workflows.Agents - Nov 16, 20260security: AI Cybersecurity Harness with Continuous Agents0security is an open-source AI cybersecurity harness that enables the creation and deployment of continuous security agents. These agents are designed to identify, validate, fix, and monitor vulnerabilities across various components including code, binaries, dependencies, infrastructure, and AI agents themselves. The project offers a Web UI, CLI, and CI integration, operating on user-provided infrastructure.Agents - Nov 14, 2026Strands Box: Open Source AI Agent SandboxStrands Box is an open-source sandbox for AI agents, written in Rust. It restricts agent execution, file access, and network reach using OS isolation and semantic policies defined in Dogwood. The system supports macOS on Apple silicon and Linux ARM64, with experimental x86-64 support.AI Coding - Nov 15, 2026OpenCode Mobile: Run AI Coding Agents on AndroidOpenCode Mobile is a new Android application that allows users to run various AI coding agents, including Claude Code and OpenCode, directly on their phones without needing a computer. The app provides a Linux environment, chat interface, diff viewer, and Git integration, enabling users to manage and steer agents remotely or locally.Agents - Nov 16, 2026Floe Agent: Open-Source AI Workspace for iOS and AndroidFloe Agent is an open-source AI workspace designed for iOS, iPadOS, and Android, offering an on-device Linux environment. It supports BYOK (Bring Your Own Key) for cloud AI models and local model inference on supported iOS devices, enabling users to work with documents, code, and creative projects locally.