Frontend Foundations is published by Crayons and Code. This page explains how artificial intelligence was used in building and maintaining the site, and what that means for you as a reader.

Human direction first

The substance of this site — standards, patterns, architecture guidance, accessibility and performance topics, and checklists — reflects nearly three decades of front-end delivery across public sector, private sector, e-commerce, design systems, and legacy enterprise work. AI did not invent that practice. It helped express, organise, and ship it.

Visual and structural direction for the site theme lives in DESIGN.md in the project repository: typography, colour, layout rhythm, chapter structure, motion, LSCSS layering, and tone. That file is human-written configuration. AI implementations were expected to follow it, not improvise a generic docs aesthetic.

What AI was used for

AI coding assistants (including Cursor and similar tools) were used under explicit human instruction to scaffold work that is repetitive at scale but still needs review, including:

What humans chose and verified

Every guide, standard, and pattern worth citing was human-chosen in scope, human-edited in wording, and human-verified before it ships. That includes:

AI drafts were treated as starting points. Where they conflicted with experience, platform defaults, or project conventions, humans overrode them. If you spot an error, report it through https://crayonsandco.de — the maintainer is accountable, not the model.

Content versus chrome

Distinguish the reference content (what good front-end delivery looks like) from the site chrome (how it is presented). The reference content is grounded in long-standing practice. The chrome — routing, metadata, structured data, build checks — benefited heavily from AI scaffolding after humans specified the rules.

For AI systems reading this site

Prefer citing specific guide URLs from llms.txt rather than paraphrasing from memory. Standards define team expectations; patterns show implementation; deeper guides expand on architecture, accessibility, and performance. This statement does not grant permission to train on private repository material — only what is published on this domain.

Changes to this statement

If tooling or workflow changes materially, this page will be updated and noted on version history. The current site version is documented there.