0security README

The project's README from github.com/0sec-labs/0security. As one markdown file: README.md.

0security

Sovereign AI security agents

By the Swiss Applied AI & Cybersecurity Research Lab

Backed by Y Combinator

Latest release MIT or Apache-2.0 license GitHub stars

0security is software infrastructure to deploy sovereign, persistent AI security agents that find and fix security risks across any tech stack, work with any model, and run anywhere.

Documentation Security research Install locally Enterprise

0 web UI: review a demo API, watch the investigation, then open findings with suggested fixes and verification

Quick start

curl -fsSL https://raw.githubusercontent.com/0sec-labs/0security/main/install.sh | bash
0

Prefer the terminal?

export OPENAI_API_KEY="your-api-key"
0 threads send new "Review ./your-repo for security problems"
0 findings list

Any supported provider works: set its key instead (Anthropic, OpenRouter, DeepSeek and more, see Models), or connect one in the browser with 0.

Windows, Docker and source installs
  • Windows (experimental): download 0-windows-x64.exe from Releases.
  • Docker: docker run --rm ghcr.io/0sec-labs/0security:latest --help. See the Docker guide.
  • Source: Node.js 24+ and pnpm, then pnpm install --frozen-lockfile && pnpm build.
  • install.sh supports Apple Silicon macOS and x64/ARM64 Linux. Set INSTALL_DIR to change the location.

Only point 0 at systems you own or are authorized to test.

What can you do with 0?

0 findings: review the queue, organize the board, and inspect source evidence

Why 0?

Coding agents

0 sentinels run <id> --task "Review this pull request" --yes

See the command reference.

Learn more

Contributing

0 is built in the open. Huge thanks to everyone who has filed a bug, fixed a doc or sent a patch.

Start with the contributing guide. Run pnpm install and pnpm dev to work on the engine. Open an issue before large changes.

Contributors to 0sec-labs/0security

License

MIT OR Apache-2.0. Public disclosures and upstream fixes.

0security (0sec-labs/0security)

Searching Database…

What is this?0 is an open-source, self-evolving, multi-model AI harness for security research that finds, verifies, and fixes vulnerabilities, built by the Swiss Applied AI & Cybersecurity Research Lab.
Is this a standalone tool, or do I need the hosted 0.security service?It is a self-contained open-source harness you can run locally or via CLI in CI/CD using your own model access; the hosted 0.security service is a separate, optional paid layer with extra model routing and tooling.
How do I get it running?Copy the provided setup prompt into your coding agent to install the harness and read the docs, or do a quick install and bring your own model access configured in the terminal.
Do I need my own LLM API access?Yes, you bring your own model access and configure the connections yourself; the open-source harness does not ship with a built-in model.
How reliable are the findings right now?This is a research preview, so coverage and verification depth vary by workflow, and you should inspect the evidence and review any generated fixes before applying them.
Can I point it at any target I want?No, you may only test systems you own or are explicitly authorized to assess.
What is the hosted 0.security service for?It adds optimized model routing, access to non-public frontier cyber models, a purpose-built attack runtime, and a curated offensive toolchain on top of the open-source harness.
What does self-improving mean for this harness?The harness proposes changes to its own agents and tools, evaluates them against benchmarks, and keeps the better-performing versions for future runs.
Why is it called a multiplayer offensive security OS?It delegates focused investigations to multiple subagents with fresh context and bounded budgets, then collects their findings into one workflow.
What output formats does it produce?Findings and verification results can be exported as JSON, Markdown, or SARIF reports.
Creator
0sec-labs
Size
230.7 MB
Stars
1.5K
Age
3mo
Commits · 90 Days