Skip to content
Sign in
Agentic Design Fundamentals

Run the designer–agent loop on real work — and keep what must stay human.

The foundation course for the whole curriculum. It covers what actually changed when capable coding agents, MCP, and design-as-code formats arrived together, the designer-agent loop that replaces tool-centric production, how to brief and harness an agent so its first draft is worth critiquing, and where the approach still fails. The material follows Part I of The Agentic Designer and the school's published briefing, harness, and critique articles, with worked examples drawn from real runs rather than demos.

This course is free with an account. No payment needed.

Modules
6 modules
Total time
≈4h 45m
Last updated
2026-10-09
Start with module 1

Curriculum

6 modules

Slides are the lesson; study guide and transcript come with each module.

Time commitment

≈4h 45m

About ≈5 hours across the full outline — work at your pace.

Access

Free with account

No payment for this course.

What you will build / 01

Artefacts you can take back to the team.

The free foundation for every paid course. Vocabulary, briefing, harnesses, and critique — before you pick a tool.

01

A plain-language explanation of agentic design your team can reuse

02

A brief that carries intent, constraints, and review criteria

03

A starter harness (instruction files, tokens, path rules) for consistent sessions

04

A critique and visual-QA habit that catches agent failure modes early

Cream paper stack with a cobalt loop of arrows, forest-green wax seal, and yellow pencil on warm paper
Curriculum / 02

Module outline

Open the first module you have not finished. The slides are the lesson. The study guide and the transcript are the same module if you would rather read.

01

The Agentic Design Paradigm

What actually changed when capable coding agents, MCP, and design-as-code formats arrived at the same time, what an agent is and is not, where the new designer-agent loop differs from the tool-centric loop, and what stays firmly human.

02

The Designer–Agent Loop

The loop from Module 1 slowed down to working speed: what each step looks like in a real session, how to treat the agent as a junior design partner rather than a vending machine, and where the loop most often breaks.

03

Briefing and Design Intent

The brief is where design intent becomes executable. This module covers the anatomy of a design brief, the difference between vague and specific prompts in measurable terms, and how research packets feed briefs that an agent can actually act on.

04

The Design Harness

Briefs carry task-specific intent; the harness carries everything durable. This module covers the instruction hierarchy, project instruction files, path-scoped rules, skills, and design tokens as agent instructions — the layer that makes output consistent across sessions and people.

05

Design-as-Code and the Tool Landscape

Why diffable design files are the foundation the rest of the workflow stands on, what the current agent platforms and connected canvases actually offer, and how to pick a starting stack without betting the team on a moving target.

06

Critique, Quality Gates, and Limits

The closing module: how to critique agent output so it improves, which checks can be made executable, the failure modes that keep recurring, and an honest account of what agents are still bad at — finishing with a ship checklist and where to go next in the curriculum.

Inside the materials / 03

Sample lesson & slide previews

Titles and diagrams from early modules — the same materials you open after checkout or sign-in.

Loop diagram of the agentic design workflow. A designer-written brief leads to an agent plan, which passes through a human review gate before the agent generates the artifact. The artifact goes through critique, then revision, and ships only after a final human decision. Labels mark brief, review gate, critique, and ship as human-led steps, and plan, generate, and revise as agent-run steps, with a dashed feedback line from revise back to the brief.Timeline trace of one designer-agent loop session across six stages. The designer writes the brief in the first ten minutes, then reviews the agent's plan at a human gate where a contradiction in the brief is flagged and resolved. The agent runs generation on a scratch branch, the result passes through a screenshot critique at a second human gate where the audit and type check run and an invented component prop is caught, the agent runs a scoped revision, and the designer makes the ship decision and records what feeds back into the harness. Each stage is annotated with what the human does and what the agent does, with approximate timestamps from a roughly thirty-minute session.Diagram of the seven lines of a design brief — situation, user job, audience and context, direction and references, system constraints, output shape, and review criteria — each shown as a labelled row paired with a note on what the agent does with that line: locating the right files, restating the job in its plan, setting tone and density, turning adjectives into layout behaviour, reusing named components instead of inventing rules, writing output to a scratch folder at the right fidelity, and running checks on its own output before reporting done. A yellow strip carries the closing plan-first instruction, and a dark strip at the bottom lists what belongs in the harness instead: tokens, component inventory, banned patterns, and working rules in DESIGN.md, AGENTS.md or CLAUDE.md, and skills.Stacked layer diagram of the agent instruction hierarchy. From top to bottom: global and user-level instructions such as a home-directory CLAUDE.md, loaded in full every session and reserved for personal preferences; the project instruction file, CLAUDE.md or AGENTS.md at the repository root, loaded in full every session and holding facts like stack, token locations, scales, and prohibitions; path-scoped rules in .claude/rules that load only when matching files are touched; skills in SKILL.md folders whose names are always visible but whose bodies load on demand for procedures and rubrics; and the task brief at the bottom, typed each session and carrying only what is new about the task. A left-hand arrow runs from broad and durable at the top to specific to this task at the bottom, and a footer notes that always-loaded layers should stay short while procedures move into skills.

The Agentic Design Paradigm · concept

What changed: three things arrived at once

The Agentic Design Paradigm · comparison

The old loop vs the new loop

The Agentic Design Paradigm · concept

What an agent is, for designers

The Agentic Design Paradigm · comparison

Four agent platforms, one comparison

The Agentic Design Paradigm · concept

Design-as-code: the artifact becomes diffable

The Agentic Design Paradigm · diagram

The agentic design loop

Audience and outcomes / 04

By the end of this course you can…

Product designers, design leads, and design-minded builders who have used chat assistants but have not yet run an agent against a real project — or who have run one and could not get repeatable quality from it.

01

Explain the agentic design paradigm in plain terms to a design team, including what stays human.

02

Run the designer-agent loop — brief, plan, review gate, generate, critique, revise, ship — on a bounded design task.

03

Write briefs that carry intent, constraints, and review criteria instead of restating the design system.

04

Set up a design harness — instruction files, path-scoped rules, tokens, and skills — that makes agent output consistent across sessions.

05

Choose a starting agent platform and design-as-code stack with a clear-eyed view of trade-offs.

06

Apply critique loops, visual QA, and quality gates that catch agent failure modes before stakeholders do.

Who it is for

This is for you if…

  • Product designers and design leads who have used chat assistants but not a project agent
  • Builders who tried an agent once and could not get repeatable quality
  • Anyone deciding which paid course to take next
Who should skip it

This is not for you if…

  • People looking for a tool certification or a general AI literacy slideshow
  • Teams that only want a vendor rollout plan with no design practice underneath
Why register now

The practice compounds. Waiting does not.

01

Every paid course assumes this vocabulary. Starting here saves rework later.

02

You can finish the first modules with only an account — no purchase required.

03

The same loop shows up in Claude Code, systems, research, and leadership courses.

Start free

Create an account and open the first module.

When you are ready for tooling depth or a workflow specialism, the Designer Track and Everything bundles sit on the same checkout paths.

Pay once. Yours to keep, with updates.

Free with an account

Sign in to startDesigner TrackEverything bundle

Buying for a team? Team seats

FAQ

Before you register

Is Fundamentals really free?

Yes. Create an account and open the modules. No payment is required for this course.

What should I take after Fundamentals?

Most designers continue with Claude Code for Designers, then a workflow course that matches their job — systems, prototyping, review, or research.

Is this a subscription?

Pay once. Yours to keep, with updates. There is no renewing membership for new purchases.

What do updates cover?

All edits and new lessons in an owned course, for the life of the course. Bundles include new courses added to that track for 12 months.

What if a course is retired?

Course retirement gives access to its replacement.

Related material / 05

The sources behind this course.

The course teaches the practice; the books and articles carry the depth, sources, and worked runs it draws on.

The Agentic Designer

The operating model for product designers, design leads, and builders who need to understand what changes when agents join design work.

Claude Code for Designers

A practical guide for designers who want to work directly with coding agents without turning it into a programming manual.

Leading a Design Organization with Agents

The three leadership courses are sold together for US$300. Designer courses and bundles don't include them, except the Everything bundle, which does. This purchase doesn't open the designer courses.

View course

Brief an Agent Like a Design Partner

A practical workflow for turning a loose design request into a reviewable agent brief — with a traced run on this site's own field-notes page, and plan-mode guidance for Claude Code, Codex CLI, OpenCode, and Antigravity CLI.

Read article

Vague Prompt vs Specific Design Prompt

A prompt teardown with the receipts: the same pricing-page request run twice — once vague, once specific — with the generated code, audit output, iteration counts, the residual failure neither prompt fixed, and the rewrite patterns that made the difference.

Read article

Visual QA With Agents

A long-form workflow for using agents to inspect screenshots, compare implementation against intent, run accessibility checks, and produce fix-ready review notes — traced against a real review of this site's own article page.

Read article

Build a Design Harness Before You Prompt

A practical tutorial for turning design-system rules, project instructions, skills, examples, and QA checks into a reusable harness that keeps agent-generated UI from drifting into generic output — including a file-by-file walkthrough of the harness behind this site.

Read article

Design Tokens Are Agent Instructions

A practical guide to turning color, type, spacing, radius, and component decisions into token files that agents can read, use, and verify.

Read article

Design Critique Loops With Agents

A practical workflow for turning agent critique from vague opinion into structured findings, evidence, severity, revision passes, and human design decisions.

Read article

Choosing Your Agent Platform: Claude Code, Codex CLI, OpenCode, and Antigravity CLI for Design Work

An honest comparison of four terminal agents for design-to-code work — Claude Code, Codex CLI, OpenCode, and Antigravity CLI — plus the IDE and cloud surfaces that are not the same product. Volatile facts were re-read on 9 October 2026. Gemini CLI is no longer the individual peer.

Read article