0 is an open-source, Apache-2.0 cybersecurity harness written in TypeScript that combines a local browser console, a command-line tool and an MCP server so developers, security engineers and AI agents can run code analysis, pentesting, vulnerability research, reverse engineering, verification and fixing against repositories and binaries using the language models the user connects.

What it is

0 is a harness rather than a single scanner: it wraps model-driven security work in a structure of workspaces, scans, automations, runs and findings. The project is developed by 0security and presented as work from the Swiss Applied AI & Cybersecurity Research Lab, with the product site at 0.security. It is distributed as a release binary for Apple Silicon macOS and x64/ARM64 Linux, as a headless container image at ghcr.io/0sec-labs/0:latest, and as a source checkout built with Node.js 24 or newer and pnpm. The repository is categorized under Developer Tools / Version Control & Collaboration and carries topics including application-security, binary-analysis, code-review, penetration-testing, vulnerability-detection and llm-agents.

The problem it addresses is the gap between an ad-hoc model session and repeatable security work. Instead of prompting a provider once and discarding the transcript, 0 retains runs, keeps source evidence next to each finding, organizes results in a queue or on a board, and lets multi-step processes be saved as automations and re-run from the CLI, from CI or from another agent. Cost is made explicit through a --cost-ceiling flag on scans, and coverage gaps are surfaced so that an empty result is not mistaken for a clean one.

Key capabilities

• A local browser console whose Connections screen accepts a model provider or a supported subscription, after which a workspace can be selected and a chat started.
• Command-line scans such as 0 scan source ./your-repo --runtime api --depth quick --cost-ceiling 2, with review of failures and incomplete coverage before results are read as clean.
• Built-in automation templates for scans, verification, fixes and research, listed with 0 automation list --templates and executed with 0 automation run --template security-research --target ... --workspace ..., plus 0 runs list for retained results.
• Findings review that keeps source evidence attached and organizes issues either in a queue or on a board.
• CI/CD integration with explicit model access, cost limits, structured reports, a severity gate, JSON and SARIF artifacts, 0 reports import, and a reusable GitHub Action.
• An MCP server started with 0 mcp-server --workflows --workspace /absolute/path/to/repo, plus plugins, hosts and skills for connecting external agents.
• An installer that verifies release checksums, honours INSTALL_DIR, INSTALL_FOXGUARD=0 and RELEASE_BASE_URL for pinning, and installs the 0 binary under ~/.0/bin.

Who uses it and how

• Security engineers performing penetration testing and vulnerability research on systems they own or are explicitly authorized to test, following the project's scope and authorization guidance.
• Developers reviewing a local repository from the terminal with a provider key set, or from the browser console when a guided workspace is preferable.
• Teams running security reviews inside their own pipelines, where the severity gate and SARIF output feed existing code review and the run's cost is bounded by the model access they grant.
• Agent builders who expose the harness to another assistant through the MCP server, using an explicitly selected host-local execution profile.
• Organizations standardizing the workflow across teams, for which the project points to an Enterprise offering and a demo booking route.

Getting started

The documented path is a single command, curl -fsSL https://raw.githubusercontent.com/0sec-labs/0/main/install.sh | bash, followed by running 0 in a new terminal to open the local browser console. Alternatives are the ghcr.io/0sec-labs/0:latest container for headless execution and a source build using corepack enable, pnpm install --frozen-lockfile and pnpm build.

How it compares

No comparable or paid products are named in the facts supplied for this entry, so 0 stands alone in this registry rather than being positioned against named alternatives. Any contrast with commercial security platforms would have to rest on details the project has not published here.

When to use it — and when not to

Adopting 0 means operating more than a binary: a model provider or subscription must be connected, API credentials supplied, and when the headless container is used the database and reports must be persisted outside the disposable container. It is a poor fit for teams that cannot send source code to a model provider, and users on Intel macOS are outside the release matrix and must fall back to a source checkout or the container, while the Windows build is explicitly experimental. Only systems owned by the operator or covered by authorization should be scanned.

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

export ANTHROPIC_API_KEY="your-api-key"
0 scan source ./your-repo --runtime api --depth quick --cost-ceiling 2

docker run --rm ghcr.io/0sec-labs/0:latest --help

git clone https://github.com/0sec-labs/0.git && cd 0
corepack enable && pnpm install --frozen-lockfile && pnpm build
node packages/cli/dist/index.js --help

0 automation list --templates
0 automation run --template security-research \
  --target /absolute/path/to/repo --workspace /absolute/path/to/repo
0 runs list

0 mcp-server --workflows --workspace /absolute/path/to/repo