Agent skill

Ontology-Driven System Builder

by sharptoolbox in sharptoolbox/ontology-driven-dev

Runs a three-step pipeline, requirement exploration, seven-model ontology YAML, then app build, on a Flask, SQLite, and React stack with sign-off gates.

MITAuto-check passedDevelopment

SKILL.md written in Chinese; this summary is our English description.

Install Ontology-Driven System Builder

skills CLI
$ npx skills add sharptoolbox/ontology-driven-dev --skill ontology-driven-dev -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install sharptoolbox/ontology-driven-dev ontology-driven-dev --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
ontology-driven-dev
GitHub stars
408
Token cost
~1.5k tokens
SKILL.md length
444 words
Files
281 (incl. references)
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

Runs a three-step pipeline, requirement exploration, seven-model ontology YAML, then app build, on a Flask, SQLite, and React stack with sign-off gates.

  • Works in 3 steps: references/本体模型业务功能开发指导书.md(核心:步骤… → references/AI原生应用技术架构设计文档.md(技术栈 / 分层 /… → references/UI-UE界面设计规范.md(配色 token / 9pt…
  • Turning a business requirement into a formal specification with sign-off gates
  • SKILL.md covers 一、适用场景与触发, 二、三阶段管线 + 人工门禁(强顺序), 三、单阶段入口(用户可指定只跑某段) and 四、固定输出约定(技能级,仅「业务域」为参数), plus 3 more sections
  • Runs Python scripts from its folder; calls npm, pip and python

What it does

Phase one turns a business requirement into a formal specification through eight strictly ordered sub-stages, covering overall understanding, business objects, functions and rules, cross-object linkage, end-to-end flows, queries and reports, roles and permissions, and an optional UI prototype. It requires explicit user confirmation before advancing past any sub-stage and never marks a document complete while an item is still unconfirmed.

Phase two turns the confirmed requirements into seven linked YAML models, covering objects, behaviors, rules, roles, processes, queries and reports, and UI, plus a manifest, checked against consistency gates such as every user-triggered behavior being referenced by a UI action. Phase three builds the runnable browser-based system on a bundled Flask, SQLite, and React/TypeScript technical base.

When your agent uses it

  • Turning a business requirement into a formal specification with sign-off gates
  • Modeling a business domain as seven linked ontology YAML files
  • Building a working system from an already-modeled ontology

Example prompts

  • “Start requirement exploration for a contract management system.”
  • “Model this approved spec into the seven ontology YAML files.”
  • “Build the running app from these ontology models.”

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

  1. references/本体模型业务功能开发指导书.md(核心:步骤 1-10、模型→实现映射总表、审批端到端、AI 对话、检查清单)
  2. references/AI原生应用技术架构设计文档.md(技术栈 / 分层 / 语义注册表 / AI 编排 / SSE / 只读 SQL 安全边界)
  3. references/UI-UE界面设计规范.md(配色 token / 9pt / 标签右对齐 / 三类界面布局 / 完整 CSS 库)

What it can do on your machine

Read from SKILL.md and the folder at commit 2e6018f. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships script files (Python, from the files we listed), which the agent can run.

    Shell commands in SKILL.md call:

    • npm
    • pip
    • python

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npm and pip, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Ontology-Driven System Builder loads about 1.5k tokens when it runs, and up to ~70k if it reads all its reference files. Until then it costs about 63 tokens; SKILL.md has 444 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~63
When it runs · the whole SKILL.md, loaded when a task matches
~1.5k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~70k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from sharptoolbox/ontology-driven-dev at commit 2e6018f, republished under its MIT licence (© sharptoolbox). 444 words, ~1,501 tokens.

Download SKILL.mdSave it as .claude/skills/ontology-driven-dev/SKILL.md (or your agent's skills folder). This skill also uses 280 other files; get the full folder from GitHub.
name
ontology-driven-dev
description
当用户要基于业务需求,通过「需求探索 → 本体建模 → 应用构建」三步法开发一套完整、可运行的本体驱动 BS(浏览器前后端) 业务系统时使用。基于七模型本体 YAML(M1/M2/M3/M5/M6/M7/MU) 与内置 code-paas 技术底座,强制每个需求探索阶段人工确认,产出严格对齐需求文档、本体模型与代码的系统。支持整体触发与单阶段入口(仅建模/仅构建)。触发词含:本体驱动、需求探索、本体建模、七模型、code-paas、AI原生应用、业务系统开发。

本体驱动系统开发技能(Ontology-Driven Dev)

将"业务需求 → 软件需求规格 → 七模型本体 YAML → 可运行 BS 系统"的完整方法论打包为可复用技能。 技术底座为内置的 code-paas(Flask + SQLite + React/TS 单体应用,含系统管理、流程引擎、工作台、本体注册表), 输出严格受四份规范约束。

路径约定(跨工具通用):本技能内所有相对路径(如 references/、reference-example/、techbase/)均以「本 SKILL.md 所在文件夹」为根目录。

  • WorkBuddy / Claude Code / Codex 等工具在加载技能时会自动解析该根目录;
  • 若某工具未自动解析,请将下文 <本技能目录> / <技能根目录> 占位符替换为本 SKILL.md 的绝对路径(例如 C:\Users\hemin\.claude\skills\ontology-driven-dev 或 ~/.workbuddy/skills/ontology-driven-dev)后执行。

一、适用场景与触发

  • 用户给出一段业务需求(一句话或一段描述),希望产出一套可运行业务系统。
  • 用户明确要求"需求探索确认 / 本体建模 / 七模型 / 基于 code-paas 构建 / AI 原生应用"。
  • 自然语言示例:「帮我开发一个 XX 管理系统」「把这段需求做成本体驱动开发」「基于这份需求规格说明书生成系统」。

二、三阶段管线 + 人工门禁(强顺序)

阶段一与阶段二、阶段三之间强顺序;阶段一内部八阶段也强顺序,且每一阶段都必须人工确认后才可推进。

阶段一:需求探索 → 软件需求规格说明书
  • 唯一依据:references/AI需求探索与确认提示词V9.0.md(其附录一即《软件需求编写规范 V9.0》全文,是本阶段格式 / 编号 / 图表 / 自检的唯一基准)。
  • 严格按该提示词的"阶段零 ~ 阶段七"八阶段推进:
    • 阶段零 总体理解确认 → 阶段一 业务对象 → 阶段二 业务功能与规则 → 阶段三 跨对象联动识别 → 阶段四 端到端协同流与审批流 → 阶段五 查询统计与固定报表 → 阶段六 角色权限 → 阶段七 UI 原型(可选,须先确认是否需要)。
  • 人工确认门禁(强制,不可跳过):每个阶段(stage0-7)结束、进入下一阶段前,必须按提示词统一的"问题 N + AI建议 + 建议理由 + 其他选项 + 快捷回复"格式提问,并硬性暂停等待用户明确确认("按AI建议" / 选项字母 / 修改意见)后才可推进。
    • A 类(行业通用)内容:AI 自动补全,标注 [AI自动补全]。
    • B 类(企业专属、猜错会有真实业务风险)内容:必须带 AI 建议提问,标注 [待确认] / [已确认]。
    • 绝不在用户未确认时私自进入下一阶段;附录 B 存在 [待确认] 则文档不得标记"完整"。
  • 终态:对照规范第十二章 47 项自检全通过、且附录 B 无 [待确认] 后,文档标记为「完整(可进入本体建模阶段)」。
  • 输出:<业务域>-需求规格说明书-V9.md(置于当前工作项目根目录)。
  • 对照范例:reference-example/合同管理需求规格说明书-V9.md(销售合同执行管理跑通实物)。
阶段二:本体建模 → 七模型 YAML
  • 唯一依据:references/ontology_modeling_framework_v9.md(七模型元文件规范 + 各模型 YAML 模板)。
  • 输入:阶段一需求文档,尤其其附录 C 七模型建模输入基线(是后续 YAML 的确定性输入,不得再做大范围业务拆分)。
  • 产物:M1 对象 / M2 行为 / M3 规则 / M5 主体 / M6 流程 / M7 查询报表 / MU UI 共七个 YAML + manifest.json,输出到当前项目的 yaml/ 目录。
  • 建模顺序建议(见指导书 §4 步骤 1):M1 → M5 角色 → M3 规则 → M2 行为 → M7 查询 → M5 权限 → M6 流程 → MU 界面。
  • 一致性门禁(强制):建模后核对——
    • 可追溯门禁:M2 triggerType=USER_ACTION 行为须被至少一个 MU 操作功能点引用;MU 引用行为须存在。
    • M7 behaviorRef ↔ M2 queryReportRef 严格一对一。
    • M6 的 roleRef / behaviorRef / subFlowRef / ruleRef 引用均须存在;SUB_FLOW_CALL 调用图无环。
    • 每个正式查询报表与唯一 M2 QUERY 行为双向一对一;联动描述中规则条件与结论不混写。
  • 输出:yaml/m1-object-model.yaml … yaml/mu-ui-model.yaml + yaml/manifest.json。
  • 对照范例:reference-example/ 下 7 个 yaml(字段语义、命名、syncTriggers / node_graph / ASCII 布局的直接参考)。
阶段三:应用构建 → 可运行 BS 系统
  • 依据(严格按序参考):
    1. references/本体模型业务功能开发指导书.md(核心:步骤 1-10、模型→实现映射总表、审批端到端、AI 对话、检查清单)
    2. references/AI原生应用技术架构设计文档.md(技术栈 / 分层 / 语义注册表 / AI 编排 / SSE / 只读 SQL 安全边界)
    3. references/UI-UE界面设计规范.md(配色 token / 9pt / 标签右对齐 / 三类界面布局 / 完整 CSS 库)
  • 技术底座:本技能内置 techbase/(即 code-paas 干净源码,已剔除 node_modules / pycache / dist / 运行时 DB)。 第一步:将 techbase/ 整体复制为当前项目根目录下的 code-app/,随后安装依赖:
    bash
    # 复制底座(保留目录结构)
    cp -r <本技能目录>/techbase/. <当前项目>/code-app/
    # (<本技能目录> = 本 SKILL.md 所在文件夹;各工具会自动解析,否则请替换为绝对路径)
    cd <当前项目>/code-app/frontend && npm install      # 还原前端依赖
    cd <当前项目>/code-app/backend  && pip install -r requirements.txt

    底座是只读基线,扩展只在 code-app/ 内进行;techbase 自带的"客户申请/查询"示例模型(models/ 下 m1/m2/m5/m6/mu)按指导书复制改造或删除,用阶段二 yaml/ 七模型取而代之并登记 manifest.json。

  • 开发顺序(指导书 §1.4 标准流水线 10 步):
    1. 写七模型 YAML 到 code-app/models/ 并登记 manifest.json;
    2. M1 → 数据库表 DDL(聚合根主表 / 子实体从表,含 5 默认字段;编号"三位前缀+四位流水号"自动生成);
    3. 数据字典由注册表自动注册;
    4. M2 行为 + M3 规则 → services/*.py(事务内;前置校验→规则校验→状态变更→syncTriggers 联动;规则引擎 simpleeval,违反给中文提示);
    5. M5 → 系统管理种子(角色 / 权限 / 资源 / 用户;接口加 @require_permission);
    6. M6 → 流程引擎(M6 activities+branches 转 node_graph;审批角色须配置且有用户绑定);
    7. MU → 菜单 + 页面 + 路由(单表 2 列 / 主从 3 列+从表表格 / 查询 3 列+结果表格分页;AggregateRootRef 跳选框、Enum/DictionaryRef 下拉框;带审批功能"保存草稿/提交"双按钮);
    8. 强制实现右侧 AI 对话框(指导书第 7 章 + 架构文档第 9 章):底座默认 ai.enabled=false 且 backend 无 ai/、sse/ 模块,本步必须补齐——system prompt 注入本体注册表、工具注册(导航/查询/行为/只读 SQL)、SSE 流式(message_start→delta→tool_call→tool_result→render_payload→message_end)、text/table/chart/action 渲染协议、动态 SQL 严格只读白名单 + 审计;
    9. 联调:登录 → 录入 → 暂存/提交 → 多级审批(通过/驳回/退回/撤回)→ 查询,全链路跑通;
    10. 验收:对照指导书附录 A 检查清单;规则违反前端有中文提示;界面符合 UI-UE 规范。
  • 输出:code-app/(可运行系统)。默认账号见 techbase/README.md(admin/admin123 等)。
Show full SKILL.md (108 more words)Show less

三、单阶段入口(用户可指定只跑某段)

  • 仅建模:用户已提供需求规格说明书 → 直接从阶段二开始,产出 yaml/ 七模型。
  • 仅构建:用户已提供七模型 YAML → 直接从阶段三开始(复制 techbase → code-app 并实现)。
  • 重确认需求:已产出需求文档但有修订 → 回到对应阶段补确认。

四、固定输出约定(技能级,仅「业务域」为参数)

产物路径 / 命名
需求文档<业务域>-需求规格说明书-V9.md(项目根)
本体模型yaml/(7 yaml + manifest.json)
业务系统code-app/
技术底座来源技能内置 techbase/(运行时复制到 code-app)

若用户显式要求其他路径/命名,以用户指定为准;否则一律采用上表。

五、关键纪律(不可违反)

  1. 人工确认不可替代:阶段一每个阶段必须硬暂停等人确认;附录 B 有 [待确认] 则文档不得标记完整。
  2. 模型是唯一语义来源:代码 / 表 / 接口 / 菜单 / 权限 / 流程 / 规则全部可回溯到某个本体模型元素,禁止"模型一套、代码一套"。
  3. 底座不动:扩展只在 code-app/ 内,techbase/code-paas 是只读基线,不就地改。
  4. 严格对齐四份规范:开发全程遵守《本体模型业务功能开发指导书》《AI 原生应用技术架构设计文档》《UI-UE 界面设计规范》《ontology_modeling_framework_v9》的强制条款(5 默认字段、跳选框、双按钮、AI 只读、逻辑删除 flag=0、状态机)。
  5. AI 对话强制:阶段三必须实现右侧 AI 对话框(含只读安全边界)。
  6. 事务与联动边界:一个行为只改一个聚合(聚合内主从同事务);跨聚合靠 syncTriggers 或流程编排。

六、参考资源索引

  • 方法论文档(5 份):references/
    • AI需求探索与确认提示词V9.0.md(含《软件需求编写规范 V9.0》全文)
    • ontology_modeling_framework_v9.md
    • 本体模型业务功能开发指导书.md
    • AI 原生应用技术架构设计文档.md
    • UI-UE界面设计规范.md
  • 黄金范例(仅文档 + yaml):reference-example/(销售合同执行管理跑通实物,对照参考)
  • 技术底座(code-paas 干净源码 + requirements.txt + README):techbase/

七、运行说明(给用户)

© sharptoolbox, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 280 other files (references) in the repository root of sharptoolbox/ontology-driven-dev.

  • SKILL.md
  • LICENSE
  • README.md
  • README_EN.md
  • code-app-example/.gitignore
  • code-app-example/README.md
  • code-app-example/backend/ai/__init__.py
  • code-app-example/backend/ai/chat.py
  • code-app-example/backend/ai/config.py
  • code-app-example/backend/ai/llm.py
  • code-app-example/backend/ai/prompt.py
  • code-app-example/backend/ai/tools.py
  • code-app-example/backend/api/__init__.py
  • code-app-example/backend/api/ai.py
  • code-app-example/backend/api/auth.py
  • code-app-example/backend/api/contract.py
  • code-app-example/backend/api/flow.py
  • … and 264 more

Open the folder on GitHubat commit 2e6018f

Compare with similar skills

Ontology-Driven System Builder next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Ontology-Driven System Builder compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Ontology-Driven System Builder this skillsharptoolbox/ontology-driven-dev408—~1.5kAutomated safety check: PassMIT
Electron Multi-Process ArchitectureiOfficeAI/AionUi33k1 repos~1.8kAutomated safety check: PassApache-2.0
Evolutionary Modular Architecturetech-leads-club/agent-skills7k—~3.7kAutomated safety check: PassCC-BY-4.0
Code Review SkillRain-kl/OpenFlare288—~2.3kAutomated safety check: NotesMIT
Audit Flowzebbern/claude-code-guide4.7k—~4.2kAutomated safety check: PassMIT
Rust Architectsergigp/yarrtube133—~6.1kAutomated safety check: PassMIT

Similar skills

  • Tells the agent where new code belongs in an Electron multi-process project and which APIs each process may use, with rules for new bridges, services, agents and workers.

    33k GitHub starsUsed in 1 repo~1.8k tokens
    DevelopmentAuto-check passed
  • Evolutionary Modular Architecture

    tech-leads-club/agent-skills

    Guides design of modular-monolith platforms with DDD, flat-by-aggregate modules, anti-corruption layers, outbox events and resilience, plus an architecture document with SVG diagrams.

    7k GitHub stars~3.7k tokensUpdated 17 days ago
    DevelopmentAuto-check passed
  • Code Review Skill

    Rain-kl/OpenFlare

    Provides comprehensive code review guidance for React 19, Vue 3, Angular 17+, Svelte 5, Rust, TypeScript, Java, PHP, Python, Django, Go, C/.NET, Kotlin, Swift, NestJS, C/C++, and more.

    288 GitHub stars~2.3k tokensUpdated today
    DevelopmentAuto-check: notes
  • Audit Flow

    zebbern/claude-code-guide

    Interactive system flow tracing across CODE, API, AUTH, DATA, NETWORK layers with SQLite persistence and Mermaid export.

    4.7k GitHub stars~4.2k tokensUpdated today
    DevelopmentAuto-check passed
  • Rust Architect

    sergigp/yarrtube

    Architecture, naming, and testing conventions guidelines for architecting Rust codebases.

    133 GitHub stars~6.1k tokensUpdated yesterday
    DevelopmentAuto-check passed
  • MVP Technical Design

    KhazP/vibe-coding-prompt-template

    Writes an MVP technical design from agreed requirements, covering architecture, data ownership, integration contracts, deployment and tradeoffs, then hands off to the next stage.

    3.1k GitHub stars~512 tokensUpdated 3 days ago
    DevelopmentAuto-check passed

Categories

Questions about Ontology-Driven System Builder

What does Ontology-Driven System Builder do?

Runs a three-step pipeline, requirement exploration, seven-model ontology YAML, then app build, on a Flask, SQLite, and React stack with sign-off gates. Phase one turns a business requirement into a formal specification through eight strictly ordered sub-stages, covering overall understanding, business objects, functions and rules, cross-object linkage, end-to-end flows, queries and reports, roles and permissions, and an optional UI prototype. It requires explicit user confirmation before advancing past any sub-stage and never marks a document complete while an item is still unconfirmed.

When should I use Ontology-Driven System Builder?

Ontology-Driven System Builder fits situations like: turning a business requirement into a formal specification with sign-off gates; modeling a business domain as seven linked ontology YAML files; building a working system from an already-modeled ontology.

How do I install Ontology-Driven System Builder in Claude Code?

Run `npx skills add sharptoolbox/ontology-driven-dev --skill ontology-driven-dev -a claude-code`. Or copy the skill folder (the sharptoolbox/ontology-driven-dev repository) into .claude/skills/ontology-driven-dev in your project. Claude Code loads it when a task matches its description.

How do I install Ontology-Driven System Builder in Codex?

Run `npx skills add sharptoolbox/ontology-driven-dev --skill ontology-driven-dev -a codex`. Or copy the skill folder (the sharptoolbox/ontology-driven-dev repository) into .agents/skills/ontology-driven-dev in your project. Codex loads it when a task matches its description.

Can I use Ontology-Driven System Builder in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add sharptoolbox/ontology-driven-dev --skill ontology-driven-dev -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/ontology-driven-dev, .gemini/skills/ontology-driven-dev, .github/skills/ontology-driven-dev and .opencode/skills/ontology-driven-dev in your project.

What does Ontology-Driven System Builder need to run?

Going by SKILL.md and its folder, Ontology-Driven System Builder needs Python for the scripts in its folder and the command-line tools its instructions call (npm, pip and python).

Does Ontology-Driven System Builder access the network?

SKILL.md contains no URLs. Its commands use npm and pip, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Ontology-Driven System Builder safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Ontology-Driven System Builder use?

Ontology-Driven System Builder is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Ontology-Driven System Builder use?

About 1.5k tokens (SKILL.md is roughly 6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 69k tokens, read only when the agent opens those files.

What are the alternatives to Ontology-Driven System Builder?

Skills that share tags, products or a category with Ontology-Driven System Builder: Electron Multi-Process Architecture (iOfficeAI/AionUi, 33k stars), Evolutionary Modular Architecture (tech-leads-club/agent-skills, 7k stars), Code Review Skill (Rain-kl/OpenFlare, 288 stars) and Audit Flow (zebbern/claude-code-guide, 4.7k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Ontology-Driven System Builder?

sharptoolbox (a GitHub user) maintains it in sharptoolbox/ontology-driven-dev, which has 408 GitHub stars. The repository was last updated on September 30, 2026.

Source: sharptoolbox/ontology-driven-dev on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.