Stack-agnostic core of the AI Unified Process (AIUP) — a structured, requirements-first workflow for taking a software project from raw vision to use case specifications.

aiup-core is the foundation plugin of the AI Unified Process. It is technology-independent: it works with any language, framework, or stack because it stops at the specification boundary. Implementation and testing are handled by stack-specific plugins (e.g. aiup/aiup-vaadin-jooq) that build on the artifacts this plugin produces.

What it does

The plugin automates the analysis phases of the AI Unified Process, adapted from the Rational Unified Process — Inception → Elaboration → Construction. Every project starts from a written vision and proceeds through requirements, an entity model, and use case specifications. Nothing gets built without a use case.

This prevents the most common failure mode of AI-assisted development: jumping straight to code from a vague prompt and producing something that half-works and can't be maintained.

Skills

Each skill is also available as a slash command. Skills pick up where the previous one left off by reading the files written along the way, so you can inspect or edit any artifact before continuing.

Phase	Skill / command	Description
Inception	`/requirements`	Generate a structured requirements catalog (user stories, NFRs, constraints) from `docs/vision.md`
Elaboration	`/entity-model`	Create an entity model with a Mermaid ER diagram and attribute tables
Elaboration	`/use-case-diagram`	Generate a PlantUML use case diagram mapping actors to use cases
Construction	`/use-case-spec`	Write detailed use case specifications (flows, pre/postconditions, rules)
Any	`/reverse-engineer`	Recover use case diagram, use case specs, and entity model from existing code

Workflow

Inception          Elaboration                          Construction
─────────────────  ──────────────────────────────────   ─────────────────
/requirements  →  /entity-model  →  /use-case-diagram  →  /use-case-spec

The skills produce and consume a set of artifacts under docs/:

docs/vision.md — input you provide (product vision, target users, goals)
docs/requirements.md
docs/entity_model.md
docs/use_cases.puml
docs/use_cases/UC-*.md

Inheriting a legacy codebase? Start with /reverse-engineer — it walks the existing code, configuration, and schema and produces the same docs/use_cases.puml, docs/use_cases/UC-*.md, and docs/entity_model.md artifacts the forward workflow would have produced, giving you a documented baseline to work from.

MCP servers

Server	Purpose
context7	Fetches current library/framework documentation on demand during analysis

Installation

Install from the Tessl registry:

tessl install aiup/aiup-core

Prerequisites

A docs/vision.md file at the root of your project describing the product vision, target users, and high-level goals. The /requirements skill reads this file to derive your requirements catalog — the richer it is, the better the results.

Next step

Once your use case specifications exist, add a stack-specific plugin to implement and test them — for example aiup/aiup-vaadin-jooq for the Vaadin + jOOQ stack.

aiup/aiup-core

README.md

aiup-core