NullSquare

concept

Welcome to NullSquare

What NullSquare is, how it is organized (organization → scope → target → run → outputs), and who does what.

NullSquare is an autonomous offensive-security platform. You point it at an environment you are authorized to test, say what you care about, and an AI agent plans the work, runs it inside the boundaries you set, and hands back assets, validated findings, evidence, and a report — the output of a senior pentest, on demand.

This is the one-page map: how the platform is organized and who does what. Everything else lives in the app, where the assistant has your current context.

What you will learn

  • The operating model. Organization → scope → target → run → outputs.
  • Who does what. The agent does the security work; you own authorization and the decisions.
  • Three products. Pentest, Compliance, and Code Review on one platform.

How it is organized

Five objects make up the platform. Once they click, every screen falls into place.

Everything flows left to right: an organization holds scopes, a scope authorizes targets, a run assesses them, and the outputs accumulate over time.
  • Organization — your account: team, roles, billing, and integrations.
  • Scope — one authorized environment. Its targets, engagement brief, credentials, and runners all live inside it.
  • Target — a domain, host, or internal range you authorize the agent to assess.
  • Run — one goal-driven assessment. The agent plans it, executes it, and produces the output.
  • Outputs — the durable record: assets, findings, evidence, and reports, tracked over time.

Who does what

NullSquare is autonomous, not unsupervised. The agent does the security work; your team owns the boundary and the decisions that follow.

  • The agent — plans the work, maps the reachable surface, tests inside the rules, preserves evidence, and writes the report.
  • You — authorize the scope and targets, add the context that tells it what matters, and triage the findings.

Scope is the boundary — and adding a target is a claim

The agent never touches anything outside an active scope. But adding a target says you own it, or have written permission to test it. Only add environments you are authorized to assess.

Where to go next

New here? The Quickstart takes you from signup to a first result in minutes. Otherwise, jump to the product you came for — Pentest Agent, Compliance Agent, or Code Review. Each is a single page that walks its whole flow.

Related articles

Last updated Jul 14, 2026