It's a folder with a markdown file in it plus optional additional reference files and executable scripts.

The clever part is that the markdown file has a section in it like this: https://github.com/datasette/skill/blob/a63d8a2ddac9db8225ee...

  ---
  name: datasette-plugins
  description: "Writing Datasette plugins using Python and the pluggy plugin system. Use when Claude needs to: (1) Create a new Datasette plugin, (2) Implement plugin hooks like prepare_connection, register_routes, render_cell, etc., (3) Add custom SQL functions, (4) Create custom output renderers, (5) Add authentication or permissions logic, (6) Extend Datasette's UI with menus, actions, or templates, (7) Package a plugin for distribution on PyPI"
  ---

On startup Claude Code / Codex CLI etc scan all available skills folders and extract just those descriptions into the context. Then, if you ask them to do something that's covered by a skill, they read the rest of that markdown file on demand before going ahead with the task.

Any particular libraries you use to generate TF-JSON from jsonnet?

I wrote a generator a little while ago that can create jsonnet libraries from the TF schemas: https://github.com/Duologic/soysonnet

Example lib here: https://github.com/Duologic/soysonnet-aws

I only needed it for AWS so I didn't spend more time on it.

This is a bummer. I don't particularly like Pulumi but use it anyways because for my use cases being able to write actual code is really impactful. Sucks to see fewer options in that space

In the photo of him on his last day [0], there's a cassette deck on his desk.

That could be something mundane, but I'd like to believe something crazy happens if you yell at it [1]...

[0] https://www.brendangregg.com/blog/images/2025/brendanoffice2...

[1] https://www.youtube.com/watch?v=tDacjrSCeq4

I feel very comfortable with Django on the frontend, what are you missing there? I usually use Tailwind or Bulma, with HTMX and AlpineJs. I feel like the experience can be very much React like, even if you leave out HTMX. The frontend game of Django really changed about 2 years ago (at least for me).

Congrats to the Django team!

If anyone is curious, I've been maintaining a Docker Compose based Django + Celery + Postgres + Redis + esbuild + Tailwind starter app for years and just updated it for Django 6.0 at https://github.com/nickjj/docker-django-example.

The only thing I haven't done is pre-configure the new CSP settings because I want to let that marinate a bit before putting it in as a default.

Seconding this. I pivoted to tech in my early 30s and feel I've missed out on a lot of community building opportunities.

I get it. This is roughly me, I don't always have the best answers, but I know most things can always be done better. I've coined a few different terms over the years such as "marketing driven development" when I wind up working in places where the marketing team is driving the devs off a cliff, and pushing new features at the expense of ever having time to deal with technical debt. The industry really needs "Tech Debt Thursdays" or something.

There's always way more work to do and those key enhancements or research stories that could improve everything get deprioritized.

My advice for building something like this: don't get hung up on a need for vector databases and embedding.

Full text search or even grep/rg are a lot faster and cheaper to work with - no need to maintain a vector database index - and turn out to work really well if you put them in some kind of agentic tool loop.

The big benefit of semantic search was that it could handle fuzzy searching - returning results that mention dogs if someone searches for canines, for example.

Give a good LLM a search tool and it can come up with searches like "dog OR canine" on its own - and refine those queries over multiple rounds of searches.

Plus it means you don't have to solve the chunking problem!