← All skills

design

External skill

Urban Design Skills

A Claude Code plugin covering the arc of a masterplan — site analysis, blocks and density, street sections, TOD, zoning and certification scoring — with Python calculators that compute density, FAR, walkability and parking rather than estimating them.

EXTERNAL SKILL

Licence · MIT

Created and maintained outside ArchitectureLM. Listed here with credit to its author and subject to its own licence. We do not host the files — the link goes to the original repository.

https://github.com/Abhinavbwj/Urban-Design-Skills-Claude

https://github.com/Abhinavbwj

Most AI assistants discuss urban design from whatever they absorbed in training. This plugin replaces that with a stated body of knowledge: theorists, quantitative standards, certification frameworks, and a set of Python calculators that do the arithmetic rather than estimating it.

By the author's own count it runs to 18 interconnected skills and 7 calculators, drawing on 40-plus theorists and six sustainability certification systems. In practice it covers the arc of a masterplan — site analysis, block and density, street sections to NACTO dimensions, public space against Gehl's criteria, TOD to the ITDP standard, zoning and form-based codes, climate response, cost, and a scorecard at the end.

Why it matters here. The calculators are the interesting part. Density, FAR, walkability, parking, green space and block optimisation are computed by scripts with declared assumptions and defaults, not produced as plausible-sounding numbers. When a figure lands in a design and access statement, that distinction is the whole of it — you can show where it came from.

What it costs you. It is a Claude Code plugin, so it needs a terminal and a local install. The standards are largely North American and European: NACTO, ITDP, LEED-ND, BREEAM, with Estidama included. Check them against your jurisdiction before quoting anything to a planning authority. And the outputs are still a starting position for your judgement, not a substitute for it — the numbers are computed, but the assumptions behind them are the author's defaults until you change them.

This is an external resource. Created and maintained by Abhinav Bhardwaj under the MIT licence, and not hosted, modified or supported by ArchitectureLM.

VERSION

main

UPDATED

September 2, 2026

FORMAT

workflow

Newsletter

Five minutes a week on what actually changed.

Subscribe

PUBLICATION

ArchitectureLM

SUBJECT

AI in architecture

NEWSLETTER

Weekly, free

© 2026 ArchitectureLM. All rights reserved.