Trajectory

Trajectory is an explainable project-planning and decision-support engine that recommends the best next task from a project's current state.

Portfolio Focus

The project is designed to demonstrate:

• dependency-graph algorithms
• deterministic decision systems
• critical-path analysis
• explainable scoring
• scenario simulation
• domain/application/infrastructure separation
• rigorous automated testing
• measured performance
• polished product UX
• disciplined agent-driven development

Repository Guide

• PROJECT_PLAN.md — product and engineering specification
• AGENTS.md — repository-wide OpenCode instructions
• docs/case-study.md — architecture case study (algorithms, design, performance, testing)
• docs/tasks.md — implementation backlog
• docs/handoff.md — current continuation state
• docs/progress.md — append-only history
• docs/decisions.md — architectural/domain decisions
• docs/architecture.md — architecture

Status

Domain core complete (model, dependency graph, scheduling and critical path, decision engine, scenario simulation). Persistence, application services, a React/Vite UI (recommendation dashboard, project/task/dependency editor, graph visualization, scenario comparison), integration coverage, and algorithm benchmarks are implemented and passing.

Core architecture, decision-engine weights, scheduling/calendar semantics, and final product naming remain under human review. Do not treat provisional product decisions as final until they are reviewed.

Engineering Principle

The deterministic domain engine is the source of truth. Any future AI layer must not arbitrarily determine project state or replace deterministic calculations.