ontology-driven-dev skill
Builds a business system from a requirements text: discovery, seven ontology YAML models, then an app with manual sign-off.
- 547 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The agent generates and runs app code and installs npm, pip or Go dependencies.
- The built-in AI chat sends schema and queries to the chosen model; the key is stored in the app.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/ontology-driven-devDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The agent generates and runs app code and installs npm, pip or Go dependencies.
- The built-in AI chat sends schema and queries to the chosen model; the key is stored in the app.
This is third-party code. Review the repository files before installing.
What it does
The `ontology-driven-dev` skill turns a business description into a working web system in three steps: requirements discovery, ontology modelling, app build. The model is seven YAML files covering objects, processes, rules, reports, UI and more, from which database tables, screens, menus, permissions and approval flows are generated. Each of the eight discovery stages pauses for human confirmation. A `code-paas` technical base ships in two variants: Flask with SQLite and React, or Go with Gin and PostgreSQL. The generated app gets a built-in AI chat that can only read the database. A sales-contract management example is included.
Who it is for. Analysts and internal-system developers who need requirements, data models and code to follow from one another.
Good fit when
- A contract approval note should become a working system with roles and reports.
- The client changes requirements and models and code must change together without drift.
- The team wants hard human sign-off points at every discovery step.
- You need a starter technical base with workflows, permissions and a model registry.
- The finished app needs a built-in AI assistant with safe database reads.
Not a fit when
- You want a weekend prototype without formal requirements.
- You need a mobile client or library: the skill produces web apps only.
- The team will not pause for sign-offs: the skill does not advance without them.
Example request
Use ontology-driven-dev: explore requirements for a sales contract tracking system and stop for confirmation after each stage.Limitations
The methodology and references are in Chinese and the English README is shorter. Node.js, Python or Go and dependency installs are needed, so the sandbox requires network. The skill imposes the single `code-paas` technical stack. Every discovery stage waits for confirmation and cannot be skipped. In the example the AI key is cleared and you connect your own OpenAI-compatible model in the app UI. For Codex and Cursor the skill is wired in by hand through `AGENTS.md` or rules.
How to disable. Delete `~/.claude/skills/ontology-driven-dev` or the project `.claude/skills/ontology-driven-dev` and restart the agent. Generated apps and YAML models stay in the project.
FAQ
What are the seven models?
A set of YAML files M1, M2, M3, M5, M6, M7 and MU for objects, processes, rules, reports and UI, from which tables, menus, permissions and flows are generated.
Can I skip the confirmations?
No. At each of the eight discovery stages the agent stops and waits for a human answer.
Which stacks are included?
A Python version on Flask, SQLite and React, and a Go version on Gin, PostgreSQL, goose and React 19.
Does it work outside Claude Code?
Yes. The README covers Codex, Cursor, Cline and Aider. Codex wires it through `AGENTS.md` and Cursor through project rules.
What about the in-app AI assistant?
The generated app gets a side chat with tool calls, and the database is read through read-only queries. You enter the model key in settings.
Discussion
No questions or reviews yet.
Related
SkillsA WorkBuddy skill that turns a business description into an ontology and generates a new skill with SQLite, an API and forms.
A set of 21 skills for Odoo addon development, OCA migration, reports, web security and architecture review.
73 skills for a delivery pipeline: plan, code, review and QA, ending in a ready, labeled PR in your own repository.
The SBTD skill set links specs, behavior, tests and domain language so an agent works through modes and gates.
Skills for spec-driven development: brainstorming, planning, debate-style review and idea grilling before the first commit.
An evidence-first development skill: the agent wraps code in a spec and a gauntlet of checks so line-by-line review becomes optional
README in short
The README explains purpose, repository layout, install for WorkBuddy, Claude Code, Codex and Cursor, and the order of the three stages. A golden contract example is included.