pmlaowangba-lab/laowangba-pmprototype-skill — explained in plain English
Analysis updated 2026-05-18
Turn a product brief into a structured information architecture before designing pages
Generate an editable Figma prototype for a B2B or consumer-facing product idea
Communicate page flow and layout intent to a design team without a finished UI
| pmlaowangba-lab/laowangba-pmprototype-skill | 1ove9/antenna-forge | ali-vilab/diffusionopd | |
|---|---|---|---|
| Stars | 64 | 64 | 64 |
| Language | — | Python | Python |
| Setup difficulty | moderate | hard | hard |
| Complexity | 3/5 | 5/5 | 5/5 |
| Audience | pm founder | researcher | researcher |
Figures from each repo's GitHub metadata at analysis time.
Requires a Codex-compatible agent environment and, for Figma output, a configured Figma MCP connection or plugin.
Laowangba PM Prototype Skill is a Codex Skill, a packaged set of instructions that teaches an AI coding agent a specific structured workflow, in this case for generating editable product prototypes in Figma, the popular design tool. It is aimed at product managers who need to turn a rough idea into a page-by-page prototype without designing everything by hand. The skill breaks prototype generation into a fixed sequence of steps rather than letting the agent freestyle. First it derives an information architecture from the initial brief, working out the entities, user roles, task flows, and pages involved, instead of applying a generic fixed list of admin pages. Next it settles a design plan covering subject, tone, design tokens, layout, and what makes the design distinctive. Before anything is generated in Figma, an anti-slop review step checks whether the proposed design has fallen into common, generic AI-generated visual patterns. The page specifications are then written out as structured YAML files, which later incremental edits can update directly. Finally the agent builds the actual editable Figma frames in batches, organized by page region. Using it requires a Codex or similar agent environment that supports loading skills, and if the user wants the agent to actually write into Figma, a configured Figma MCP connection or Figma plugin with specific generation capabilities. The skill also builds on a separate frontend-design skill for its visual design judgment, if that companion skill is not installed, the repository's own design and gate-checking documents cover the same ground. The README includes three worked examples: a membership system's customer-facing pages, an inventory management backend, and an AI price-comparison storefront, each showing the information architecture, design plan, and resulting page specifications. The author is explicit that the output is a prototype, not a finished UI design ready for delivery to a client or for a designer to build from directly, it is meant for communicating requirements, information architecture, and page flow. The project is released under the MIT license.
A Codex Skill that guides an AI agent through a structured workflow for turning a product brief into an editable Figma prototype, step by step from information architecture to design.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly pm founder.
This repo across BitVibe Labs
double-check against the repo, no cap.