What is this?
A skill for AI coding agents (WorkBuddy / Claude Code / Codex) that turns a plain-language business requirement into a runnable browser-based system through three steps: requirement exploration, ontology modeling into seven YAML models, and application construction on a built-in Flask+React tech base called code-paas.
A skill for AI coding agents (WorkBuddy / Claude Code / Codex) that turns a plain-language business requirement into a runnable browser-based system through three steps: requirement exploration, ontology modeling into seven YAML models, and application construction on a built-in Flask+React tech base called code-paas.Is this a standalone app, or a skill file that guides an AI agent?It is a skill (a SKILL.md instruction file plus reference docs and a tech base) that an AI coding agent follows step by step; you don't run it directly, it directs the agent to build an app for you.
It is a skill (a SKILL.md instruction file plus reference docs and a tech base) that an AI coding agent follows step by step; you don't run it directly, it directs the agent to build an app for you.How do I add this skill to my AI coding agent?For Claude Code, drop the folder into .claude/skills/ so it auto-discovers SKILL.md; for Codex, place it under .codex/skills/ and reference its path from codex.md or AGENTS.md; for Cursor/Cline/Aider, load or paste SKILL.md as a project rule or system instruction.
For Claude Code, drop the folder into .claude/skills/ so it auto-discovers SKILL.md; for Codex, place it under .codex/skills/ and reference its path from codex.md or AGENTS.md; for Cursor/Cline/Aider, load or paste SKILL.md as a project rule or system instruction.What do I need to actually run the generated system?Python 3.10+ for the Flask+SQLite backend and Node.js 18+ for the React+Vite frontend; the first run needs internet access to execute npm install and pip install.
Python 3.10+ for the Flask+SQLite backend and Node.js 18+ for the React+Vite frontend; the first run needs internet access to execute npm install and pip install.Can I skip the confirmation steps to speed things up?No — each of the eight requirement-exploration stages must pause and wait for explicit human confirmation before the agent can continue, and this gate is described as mandatory and unskippable.
No — each of the eight requirement-exploration stages must pause and wait for explicit human confirmation before the agent can continue, and this gate is described as mandatory and unskippable.Which AI coding agents does this actually work with?It runs in WorkBuddy, Claude Code, Codex, Cursor, Cline, and Aider, since it has no WorkBuddy-specific dependency and only needs the host tool to load the SKILL.md instructions.
It runs in WorkBuddy, Claude Code, Codex, Cursor, Cline, and Aider, since it has no WorkBuddy-specific dependency and only needs the host tool to load the SKILL.md instructions.What does the example application demonstrate?code-app-example is a full sales-contract execution management system with contract registration and tiered approval, invoicing approval, payment write-off, cross-object status linking, fixed reports, and the built-in AI assistant, complete with seed SQLite data and a default admin/admin123 account.
code-app-example is a full sales-contract execution management system with contract registration and tiered approval, invoicing approval, payment write-off, cross-object status linking, fixed reports, and the built-in AI assistant, complete with seed SQLite data and a default admin/admin123 account.What is the code-paas tech base used in step three?It's the built-in Flask+SQLite+React/TS monolith shipped with system administration, a workflow engine, a workbench, and an ontology registry — you copy it into your project and extend it instead of building the plumbing from scratch.
It's the built-in Flask+SQLite+React/TS monolith shipped with system administration, a workflow engine, a workbench, and an ontology registry — you copy it into your project and extend it instead of building the plumbing from scratch.Does the generated app come with AI chat built in?Yes, every generated system has a mandatory AI chat panel fed by the ontology registry, with tool-calling, SSE streaming, and a read-only SQL safety boundary.
Yes, every generated system has a mandatory AI chat panel fed by the ontology registry, with tool-calling, SSE streaming, and a read-only SQL safety boundary.What are the seven ontology models used for modeling?M1 objects, M2 behaviors, M3 rules, M5 actors, M6 processes, M7 queries/reports, and MU UI — these YAML files plus a manifest.json are the single source of truth that generates the database schema, APIs, menus, permissions and workflow.
M1 objects, M2 behaviors, M3 rules, M5 actors, M6 processes, M7 queries/reports, and MU UI — these YAML files plus a manifest.json are the single source of truth that generates the database schema, APIs, menus, permissions and workflow.What are similar projects?threejs-architecture-effects, JeecgBoot, cc-wf-studio, herk-edit-kit, nocobase, hand-drawn-styles, multi-agent-discuss, datapages
threejs-architecture-effects, JeecgBoot, cc-wf-studio, herk-edit-kit, nocobase, hand-drawn-styles, multi-agent-discuss, datapagesWhich topics is it in?cline, sse, flask, low-code, yaml, claude-code-skill, code-generation, vite, workflow-automation, cursor, sqlite, html
cline, sse, flask, low-code, yaml, claude-code-skill, code-generation, vite, workflow-automation, cursor, sqlite, html