Topics

Agent loops

Week 1

Every coding agent is the same small machine: a model, a set of tools, and a loop that runs until the job is done

Agentic coding tool catalogue

A tour of the agents, harnesses, MCP servers and other tools worth knowing about

AI use and academic integrity

Using AI tools in this course is allowed and expected, for students and convenors/tutors alike — this page gives a bit more detail on what that means in practice

Assessment

Week 1Week 2Week 5

The model behind the three assignments and how everything is marked: the shared rubric, what you submit, deadlines, extensions, and appeals

Backpressure

Week 3Week 5Week 6

Most of harness engineering is giving good negative feedback: the model generates, and the harness rejects what's wrong, fast and automatically

Claude Code cheatsheet

Week 2Week 3Week 4

CLAUDE.md, /clear, plan mode, and the rest of the Claude Code habits the course leans on, all on one page

Code of conduct

How we treat each other in this course

Communication

Week 1

Where to go for what, and what each channel is actually for

Context engineering

Week 4

The context window is the agent's working memory. Keep one coherent task in it, load what the task needs, and start clean when the work changes

This course is an experiment

Week 1

COMP4020/8020 is running for the first time, so here's how we'll keep you in the loop, and keep things fair, as it evolves

Course overview

Week 1

Learning outcomes, structure, and expectations

Crit agents

Week 1

The named agents, one per crit group, that sit the same coursework you do and remember what they built with you

Enrolment

Permission codes, crit-group registration, late enrolments, pre-requisites, the census date, and joining after week 1

Ethics

Should this course even exist? And should you take it?

Evaluating agent output

Week 5

The agent says it's done. Verifying that claim means running the thing, and keeping enough understanding of it to answer for what shipped

Evidence of practice

Week 6

Reconstruct one consequential decision from the transcript and the git log, diagnose the habit behind it, and change the practice from evidence rather than memory

Following along from outside

The coursework is public end to end. Here's what anyone, or anything, outside the enrolled cohort can and can't do with it

Generating images

How to generate images for your prototypes --- with the course key, or a third-party text-to-image API

GitHub Actions basics

Week 6

CI is one idea — a check that runs the moment you push, on a machine that isn't yours. GitHub Actions is the instance you happen to be using

GitHub Pages deployment

Week 1Week 3

You deploy from the first crit onwards, and GitHub Pages is how: push a commit, and a URL goes live

Your hosting access

Every full-stack repo deploys to its own app on Fly.io — this page is the whole story of the token, the deploy, and the shape your app has to keep

How browsers work

Week 2

Request, response, parse, render: what actually happens between typing a URL and pixels on screen

Interactive pages

Scrollytelling and playful interactivity in the neal.fun and pudding.cool genre, which is about as far as static goes

The lecture-demos repo

Week 1Week 2

The public repo the live demo runs in. Clone it in week 1, pull to follow along, and reset at the start of each lecture

The lecture structure

Week 1

How a lecture runs: one live demo launched on the premise you set and left running, two check-ins where you push it further, and Q&A to close

Your LLM access

Week 1Week 4

The course issues you a Claude API key. This page covers wiring it into Claude Code, what gets logged, and how your usage allocation works

Modern web best practices

Give the agent current guidance, then verify the running artefact with the browser, focused checks and specialised sensors when they earn their setup

Onboarding

Week 1Week 6

The whole course setup as a terse checklist: accounts, tools, your API key, and the course skills plugin, each linking to the full how-to

Your participation

Week 1

Everything the course asks of you, as one list — the week-1 setup, the weekly rhythm, the assignment deadlines and the standing responsibilities — with links to the detail

Skills

Week 4Week 5

A skill packages instructions and scripts once, so the agent can load it on demand instead of you explaining the same thing every session

Software and platforms

Week 1

What each platform is, who pays for what, and the course policies that come with working agentically on ANU infrastructure

Spec-driven development

Week 2

The brief poses the problem and the spec is its fixed contract: interrogate both with the agent, plan a response, then turn checkable clauses into tests

Static site generators

Week 3

A static site generator is the build step between the content you write and the pages you ship

The studio crit model

Week 1Week 6

How a studio crit works and how it's marked: the five beats of the session, the feedback protocol, and the two marks each week carries

Task tracking

Week 2Week 3Week 4

A structured task list, held outside the agent's context window, is the plan's external source of truth

What's yours in the work

Week 2Week 5

When the agent writes the code, what's left is your taste, your standards and your judgement. Every deliverable asks you to put it in writing