# Our agent skills for Wagtail developers

> Source: <https://wagtail.org/blog/our-agent-skills-for-wagtail-developers/>
> Published: 2026-10-08 00:00:00+00:00

# Our agent skills for Wagtail developers

## Agent-friendly documentation for your Wagtail tasks

We have been experimenting with agent skills [for a little while](https://wagtail.org/blog/an-agent-skill-to-upgrade-your-wagtail-site/), and here we are, with five official skills ready for all sorts of Wagtail tasks!

## Our official skills

We decided to start with the following:

- **API** : configuration, usage of[our new API](https://wagtail.org/blog/cms-with-ai-not-ai-cms-wagtail-80s-new-api/) to automate common tasks.
- **Backend** : to make the most of Django + Wagtail as a framework.
- **Content modeling** : to structure data properly with our page tree and StreamField blocks.
- **Frontend** : for page rendering, rich text, images, accessibility, SEO, and more.
- **Upgrades** : to help with automation for those critical points in a project.

All the skills focus on the expert knowledge needed to complete tasks, not the way specific tasks should be done, we leave that to you.

## Quality-controlled for real-world usage

All skills have been tested and optimized across a wide range of models, to make sure they’re relevant for all. On average, **we see tasks complete on the order of 30% faster with the skills compared to without**.

Here is an excerpt from our upcoming benchmark, showing performance across models.

Following the response we received in our [DX with AI survey](https://wagtail.org/blog/2026-ai-dx-survey/), we want to focus on a minimal set of skills that decidedly help our users get the most out of Wagtail.

## How to use them

We have spent a lot of time tuning this too, so you get the best results on Wagtail tasks without having to manage a plethora of skills or having them pollute context on other work.

**Our recommended approach** is to install one evergreen Wagtail skill plus the [Wagtail CLI](https://github.com/wagtail/wagtail-cli), and it will automatically load relevant information from pre-bundled skills files and live developer docs as needed. Install the CLI first, then install the one skill. Here’s an example with uv and a location compatible with harnesses that follow the [Agent Skills](https://agentskills.io/home) specification:

```
uv tool install wagtail-cli

# Global skill installation:
mkdir -p ~/.agents/skills/wagtail && wt skills get wagtail > ~/.agents/skills/wagtail/SKILL.md
# Local skill installation:
mkdir -p .agents/skills/wagtail && wt skills get wagtail > .agents/skills/wagtail/SKILL.md
```

It’s just the one short skills file, that will only trigger if Wagtail is involved, and will then help the agent decide whether to load the specialized Wagtail backend skill, or frontend, or both. It’s highly optimized to de-clutter your agent’s context.

There are a lot more options too - check out our [wagtail/agent-skills](https://github.com/wagtail/agent-skills) repository!

## Future skills

We’ll want to publish a few more for specific tasks but very opt-in. And we’ll want to unpublish some of our skills as we see them becoming less relevant with newer generations of models. [Feedback welcome](https://github.com/wagtail/wagtail/discussions/14438)!

And if you’re interested in how we benchmark AI models and harnesses with Wagtail and Django, be sure to sign up for [Wagtail Space 2026](https://wagtail.org/wagtail-space-2026/) where all will be revealed in November ✨