Intelligence · Strategy

Strategy & adversarial review

Strategy in OpenLegal is a structured argument, not a suggestion engine. Specialist roles debate and red-team; options, war games, and contingencies emerge — and a human holds the final call.

The strategy pipeline

DEBATE

Council synthesis

Specialist roles — each representing a perspective — debate the matter and synthesize a position.

↬
RED-TEAM

Adversarial review

An adversarial pass attacks the synthesis, exposing assumptions and weak spots before they reach you.

↬
OPTIONS

Strategy work items

Options are structured, sourced, and prioritized with dependencies.

↬
WAR GAME

Play it forward

War-game scenarios test contingencies and adversarial responses before any action is chosen.

↬
APPROVE

Human gate

Strategy output lands at an approvals surface — terminal, browser, or API — and stops there.

Design principles

Second opinions by design

Debate and red-team are built into the loop, so a single-model blind spot is an explicit review target.

Contingencies, not guarantees

War games produce options and fallbacks — the system never promises that an outcome will happen.

You remain the decision

No strategy item executes autonomously. Approval is a first-class gate, not an afterthought.

Related

Research

The authority base strategy argues from.

Read

Privacy

How the gate keeps output inside your control.

Read

Roadmap

Upcoming strategy and war-game work.

Read