How it works

From guided intake to tracked execution.

The core design principle is simple: understand the user properly, plan in the right sequence, and keep the plan alive after generation.

Phase 1

Intake validation

Verify consent, completeness, and baseline sufficiency

Owner: intake-agent

Phase 2

Profile synthesis

Build structured profile and user intelligence

Owner: profile-agent

Phase 3

Risk and evaluation gate

Downgrade unsafe or low-confidence planning before deeper agents run

Owner: evaluation-agent

Phase 4

Strategy planning

Sequence domains and choose the correct intensity

Owner: strategy-agent

Phase 5

Domain planning

Generate measurable actions by domain

Owner: domain-agents

Phase 6

Compilation and tracking

Package outputs, initialize tasks, and expose review loops

Owner: compiler + tracker