Analyzes a natural-language brief and produces a ready-to-run Foldkit program scaffold (file layout, model/messages/commands, views, and tests) following Foldkit conventions. It guides the generator through tiered complexity, API cribbing, UI component choices, and verification steps to produce production-quality code.
Invoke this skill when you want a new Foldkit app scaffolded from a feature description (single-page widgets to multi-route apps). Use it to ensure generated code follows Foldkit architecture, includes tests, and passes format/lint/typecheck gates.
Best used by powerful code-generation agents and CI-integrated bots (Copilot/Codex/GPT-style with code execution capability).
A comprehensive code-generation skill for the Foldkit framework that guides an agent through a 7-phase workflow: analyze description, study reference examples, determine file organization, generate the app, run an internal review loop (up to 3 rounds with a separate reviewer persona), and explain the result. No scripts bundled — purely instructional prompt text. Extremely thorough with specific blind-spot checklists, quality exemplars, and complexity tiers.
No security concerns whatsoever — pure instructional text with no scripts, no network calls, no credentials, no destructive commands. Well-structured multi-phase prompt with clear quality gates and review loops. Niche but valuable for its target audience.