DEV Community

Abhishek Bijjargi
Abhishek Bijjargi

Posted on

Using AI Agent Skills to Simplify Wagtail Bakerydemo Maintenance

#wagtail #git #github #ai

Building an AI Agent Skill for Wagtail Bakerydemo Maintenance

What is an AI Agent Skill

An AI agent skill is a structured markdown document that gives AI
coding assistants like Claude, Cursor and GitHub Copilot contextual
knowledge about a specific codebase. Instead of the AI having to
read through your entire source code every time you ask a question,
the skill file provides a curated reference covering project
structure, common tasks, conventions and known issues.

Wagtail recently published a blog post about agent skills for
upgrading Wagtail sites. As part of my GSoC 2026 application for
the Demo Website Redesign project, I wrote one specifically for
maintaining the bakerydemo codebase.

Why the Bakerydemo Needs This

The bakerydemo serves two critical purposes for the Wagtail
project. It is the primary demonstration tool for developers and
organisations evaluating Wagtail, and it is the standard local
development environment for Wagtail contributors.

Maintaining it involves recurring tasks like updating dependencies,
managing content fixtures, modifying templates and checking
accessibility. An AI agent skill that understands the bakerydemo
structure means any contributor can get accurate, context-aware
assistance for these tasks without needing deep familiarity with
the codebase first.

What the Skill Covers

The skill I wrote covers the following areas.

Project structure — a complete map of the bakerydemo directory
tree explaining what each folder and file contains and what it is
responsible for.

Common maintenance tasks — step by step guidance for updating
dependencies, adding new content fixtures, modifying templates,
running tests and checking accessibility compliance.

Wagtail specific concepts — explanations of StreamField,
the image library, snippets and search as they are used in the
bakerydemo specifically.

Development environment — how to set up the Docker based
development environment and the simpler local setup.

Pull request guidelines — what is expected when submitting
contributions to the bakerydemo repository.

Common errors and solutions — the most frequent issues
contributors encounter and exactly how to fix them.

How I Built It

I started by spending several hours reading through the bakerydemo
codebase on GitHub. I mapped out the project structure, read through
the models, templates and fixture files, and noted down all the
Wagtail-specific patterns used throughout the project.

I then set up the full bakerydemo development environment locally
using Docker, which gave me hands-on familiarity with the setup
process and the errors that new contributors typically encounter.
I documented these errors and their solutions directly in the skill.

Finally I structured all of this knowledge into a markdown document
following the format that AI coding assistants respond to most
effectively — clear headings, concrete code examples and explicit
step-by-step instructions rather than vague general guidance.

The Skill File

The complete skill is available as a public GitHub Gist here:
GitHub Gist

How to Use It

If you are using Cursor, open the bakerydemo project and paste
the skill content into the Cursor context or reference the gist
URL directly in your prompt.

If you are using Claude, paste the skill content at the start of
your conversation before asking maintenance questions.

Example prompt:

"Using this bakerydemo maintenance skill: GitHub Gist, help me
add a new page type for a Special Offers section. What files
do I need to create or modify?"

The AI will then give you accurate, bakerydemo-specific guidance
referencing the correct file paths, Wagtail conventions and
project structure without needing to read the source code itself.

DEMO

For Claude

Demo Video

For Anti Gravity

Prompt

Implementation Plan provided by AI

Implementation Plan: Add Special Offers Page Type

Based on the prompt, we are adding a new "Special Offers" section to the existing bakerydemo repository (bakery-demo-skill-test).

User Review Required

Since this introduces a new app into the codebase, please review the plan below.

Once approved, the implementation can be executed step-by-step.

Proposed Changes

Following the SKILL.md guidelines, this involves:

  • Scaffolding a new Django app within the Wagtail structure
  • Creating templates using accessible patterns
  • Validating via test cases

Core Settings

[MODIFY] bakerydemo/settings/base.py

  • Add 'bakerydemo.special_offers' to INSTALLED_APPS
  • This ensures Wagtail detects the new models

New App: special_offers

[NEW] bakerydemo/special_offers/__init__.py

  • Empty file to declare module

[NEW] bakerydemo/special_offers/models.py

  • SpecialOfferIndexPage → parent container
  • SpecialOfferPage → individual offer page
  • Use StreamField for flexible content
  • Integrate Wagtail image library

Templates

[NEW] templates/special_offers/special_offer_index_page.html

  • Extend {% extends "base.html" %}
  • Loop through child pages
  • Use {% image %} and {% pageurl %}

[NEW] templates/special_offers/special_offer_page.html

  • Extend base.html
  • Render StreamField blocks
  • Include accessibility-friendly structure

Tests

[NEW] bakerydemo/special_offers/tests.py

  • Basic TestCase
  • Validate model instantiation
  • Ensure compatibility with:
python manage.py test bakerydemo.special_offers

Open Questions

  • Should SpecialOfferPage include:
    • discount percentage?
    • expiration date?
    • or just flexible content?

Verification Plan

Automated:

python manage.py makemigrations special_offers
python manage.py migrate
python manage.py test bakerydemo.special_offers

Accessibility Checklist:

  • Images must include meaningful alt text
  • Forms must have proper <label> tags
  • Ensure WCAG AA contrast compliance
  • Full keyboard navigation support
  • Unique and descriptive page titles

Relevance to the GSoC Project

As part of the Demo Website Redesign project I plan to work on
during GSoC 2026, I will be introducing new templates, content
fixtures and possibly new page types. Updating the agent skill
as part of this work means future contributors and maintainers
will have an accurate reference for working with the new structures
using AI assistance.

This is a small contribution now but has compounding value over
time as more contributors use AI coding tools in their workflow.

My GSoC Application

I am applying for Google Summer of Code 2026 with Wagtail for the
Demo Website Redesign project. If you are also going through the
application process feel free to reach out on the Wagtail Slack
or GitHub.

GitHub: GitHub
Dev.to: Dev.to
Wagtail participant file: Wagtail
LinkedIn: Linkedin

Top comments (0)