Richard Lovell 🇳🇿✨

is typing...

[Steve Ruiz @steveruizok]

I'm kind of iffy on "skills discourse" but I think I came up with one so good that its worth sharing. The basic ask is to "describe the UX of my application completely". In doing so, agents will discover inconsistencies, bugs, and design errors that have never come up before

[Scott Tolinski @stolinski]

I made this for Sentry Hackweek. I'll be doing a deep dive soon, but it's a personal recorder that sends my transcripts to a router that determines if the text is a simply a todo, reminder OR if it's work. If it's work, Pi agents get started in it right away, coding or just doing research.

Built using:
ESP32-S3: https://t.co/wEai0hRAV9
1100mAh Battery: https://t.co/Uxgtc1HNlU
https://t.co/ywTjOUZdqR for all agent work
https://t.co/VRhJuaJ9E5 for deterministic workflows, automations
https://t.co/PgdQbjb9o3 to help design the case
+ a ton of sandpaper to get it smooth.

Also a adafruit mic on there

[Johannes Tscharn @JohannesTscharn]

For anyone wondering what device this is:
https://t.co/RLiAuFkH5E

Thank you @steveruizok for the inspiration!

[Steve Ruiz @steveruizok]

two demos of apps I've made for these little guys

[David Haz @davidhaz]

Introducing Canvas UI, the first ever html-in-canvas component library.

Your DOM is the render target now. Real-time shaders over real, interactive UI.

24 components. React, Vue, Svelte, vanilla TS.

Free. Open source. 🧵
https://t.co/Y5ZZDNxPUX

[Meng To @MengTo]

I'm open-sourcing my Agent Skills library.

75 skills for Codex, Claude Code, Cursor, and other agents, focused on web design, landing pages, motion, WebGL, UI styles, and assets.

A few favorites:

- Video to Super Prompt
Turns a screen recording of a design, landing page, or animation into a super detailed prompt that Fable 5 can one-shot into HTML.

- HTML to Interaction Prompts
Takes an existing HTML page, like something built in Aura, and extracts prompts for sections, buttons, animations, WebGL effects, and interactions.

- Stitched Full Page Capture
Captures the entire landing page, not just the hero, so you can use the full page as a design reference.

- Daily UI Inspiration
Combines multiple skills into an agent loop that browses the web, captures great landing pages, and turns them into detailed prompt packs.

It's free. Fork them and adapt for your own workflow.

[Meng To @MengTo]

I recorded a 28-min video on how to make next-level landing pages, mobile designs and slides with DESIGN.md

[Julie Zhuo @joulee]

A recent unlock for me on AI + data analysis: think less about prompting. Think more about cooking.

See a lot of people use AI like a microwave.

They drop in one chart, one problem statement, one KPI dip, and type: “Think like a senior analyst. What should I do?”

Then they hit analyze and act surprised when what comes back is lukewarm slop.

But good analysis is not microwave work.
It’s chef work.

If you give a great chef a microwave and say “make dinner,” you should not be shocked if the result is random.

A chef needs more than that.

They need a pantry.
They need various tools.
They need to know who they’re cooking for.
They need to know whether this is Tuesday dinner or a wedding.
They need to know what was already served.
They need to taste as they go.
They need constraints.
Same with AI.

Most people give AI one slice of the situation:
“My growth is slowing. What should I do?”
“Our retention is down. What’s happening?”
“Revenue is up. Is that good?”

That is not enough.

Because a good answer depends on other context that narrows what is actually true.

For example:

What exactly is the metric?
How is it defined?
What changed recently?
Which segments matter most?
What are we optimizing for?
What happened the last time this moved?
What constraints are real?

That’s what I mean by orthogonal context (which is a fancy way of saying, context that comes at right angles. That is independent from each other.) Different kinds of context that rule things in and out.

This is why “better prompts” are overrated.
“Act like a strategic analyst” is basically: “Cook like a Michelin chef.”

The problem is not that the model is dumb.
It’s that you gave it one thing and asked it to invent the meal.

A better question is: What are the 5–7 things my best analyst would want to know before making a recommendation?

Then, answer those questions. Give your AI the pantry and tools that it needs.

[Aiden Bai @aidenybai]

Introducing Expect

Let agents test your code in a real browser

1. Run Claude Code / Codex to QA your app
2. Watch a video of every bug found
3. Fix and repeat until passing

Run as a CLI or agent skill. Fully open source

[Matt Pocock @mattpocockuk]

Here's my AI coding workflow and all the skills I'm using:

Idea -> /write-a-prd -> PRD
PRD -> /prd-to-issues -> Kanban Board
Kanban -> ralph​.sh -> Ralph Loop
Ralph Loop -> Manual QA

Links below to skills

[Tommy Geoco 🇺🇸 @designertom]

Watch it here:

https://t.co/mw6DY66iSh

[Linus ✦ Ekenstam @LinusEkenstam]

You can try it at https://t.co/EEwnGzK3vL

[Charlie Clark @charliie]

Ok, this remotion skill is pretty cool. This took me about 20 minutes.

[Cory House @housecor]

Woah, Vercel just released the React Best Practices skill. Looks super handy.

Love the idea of encapsulating all this knowledge in a single skill. And since it's a skill, it's lazy loaded.

To install:
npx add-skill vercel-labs/agent-skills

https://t.co/gfYaUnPnvd

[Zach Bruggeman @zachbruggeman]

Get the tech spec, and read all the lessons we learned along the way on our blog:

https://t.co/rYJ3CjNhna

[Richard Lovell 🇳🇿✨ @richjavalabs]

If you ask @cursor_ai to fix your build errors, it may only look at snippets of the compiler output.

TypeScript errors are global — partial logs never converge.

Fix: make it run:

`npm run build > build.log 2>&1`

and reason over the full output.

[Alex Sidorenko @asidorenko_]

Chrome DevTools MCP in Cursor

[Richard Lovell 🇳🇿✨ @richjavalabs]

@steveruizok @Context7AI I'm getting it to do reviews on my code. Seems to be doing a good job so far

[shadcn @shadcn]

📦 I just published the Registry Directory: a list of code registries you can browse and pull code and components from.

https://t.co/2ptrwe2qd8

Built into the CLI. No config required.

[Riley Brown @rileybrown]

Damn OpenAI is really going after Zapier and n8n...

They're releasing an "Agent Builder" tomorrow (apparently).

And you'll be able to add MCP tools, Guardrails and more.

For the past three years people have been making ChatGPT wrappers.

I think the next wave will be Agent Wrappers.