Diego Francisco Valenzuela Iturra

Coffee, Music and AI

[Thariq @trq212]

we heard feedback that it's hard to know if your skills are still working with new model releases

plugin evals are here to help

run `claude plugin eval init` in your plugin folder

[Gradio @Gradio]

Can you run a workflow with eleven AI pipelines without downloading 20 GB of checkpoints and installing work environs?😎🎯

Workflow1111: Introducing a brand new way to build and share your complex gen AI workflows👇

[CopilotKit🪁 @CopilotKit]

Bring your 𝙾𝚙𝚎𝚗 𝙱𝚘𝚝 to Slack

Tag a bot in a thread and it uses its own computer to complete the task end to end.

Example:

→ opens the page
→ posts a screenshot
→ writes the report
→ and shares the file

Powered by the Channels SDK.

Open source, see below👇

[ClaudeDevs @ClaudeDevs]

New in Claude Code: claude plugin eval

See what value your plugin is adding, or if it needs more work.

You can create test cases, run your plugin or skill against those test cases, score those runs, then run each case again without the plugin to see the differences.

[Charlie Marsh @charliermarsh]

By our measurements, python-build-standalone is now the fastest CPython distribution across major platforms.

It's 10% faster than Homebrew's CPython, and 18% faster than the official python:3.14 Docker image.

[Ayush S @ayushswrites]

Turns out this side-quest is becoming infra for the biggest product breakthrough we've had all year.

[Tanner Linsley @tannerlinsley]

TanStack Charts is officially Alpha! 🥳

Since my last update:
🚢 11 releases
📚 188 examples + full shadcn chart catalog
📊 Multiple axes + polar scales
🎨 Mixed SVG + Canvas marks
☀️ Drillable sunbursts
🧹 Cleaner API

v0.16.0 is out. Go break it:
https://t.co/0GKCYjI7yI

[Emil Kowalski @emilkowalski]

Inspired by @lenjaminbeonard beautiful work on https://t.co/3uzj5lL4Zc

[Atai Barkai @ataiiam]

🎉Excited for our collaboration with @AnthropicAI

Claude Managed Agents & Agents SDK are now
AG-UI compatible!

Build Fullstack agent apps with:
- Generative UI
- Shared state
- Streaming chat

Connect your agents to Slack and MS Teams with the Channels SDK, and bring them seamlessly into OpenBot.

Check out the cookbook 👇

[Atai Barkai @ataiiam]

We made adding a @Pydantic agent to any Channel (like Slack) just one prompt.

Via the AG-UI protocol.

👉 "Read https://t.co/h6bsTzGZek and help me build my first channel with a Pydantic AI agent."

Try it out

[Atai Barkai @ataiiam]

AG-UI is the Agent-User Interaction Protocol

An open, lightweight, event-based protocol that standardizes how AI agents connect to user-facing applications.

If you haven't, give it a star 🌟
https://t.co/sVFn3lNqcS

[Atai Barkai @ataiiam]

🎉Introducing the new @CopilotKit Inspector

A complete development loop for production-grade agents:

See problem → Reproduce → Fix → Improve

The Inspector:
1. Notifies you when something breaks
2. Takes you to the cause
3. And lets you reproduce stored conversations safely

Then it turns repeated patterns from real usage into skills your agent inherits!

Already used and loved by hundreds of thousands of devs.

Setup is just one prompt 👇

Get started: https://t.co/B6dIBrKk9U

[Lydia Hallie ✨ @lydiahallie]

I've been running `/claude-api prompt-audit` when upgrading projects to Fable 5.1 and it's helped so much

It goes through your setup (CLAUDE.md, skills, etc.), flags instructions the new model doesn't need anymore and proposes a diff. Also works if you don't use the API!

[Lydia Hallie ✨ @lydiahallie]

@Nateemerson Yeah that's by design, but you can scope it by adding specific paths like `/claude-api prompt-audit CLAUDE.md .claude/` to only audit your CC setup!

[ClaudeDevs @ClaudeDevs]

You can pop out any pane in the Claude Code desktop app into its own window.

Drag the diff or terminal to a second screen while Claude keeps working in the main window, then dock it back whenever you want. You can also run sessions side-by-side or stacked.

[Mike Tamir, PhD @MikeTamir]

Anthropic released commerce-agents, a reference blueprint for building shopping and merchant agents with Claude, including runnable verticals and Claude Code integration.

https://t.co/zP5zDESX6d

#MachineLearning #AI #LLM #DeepLearning #AgenticAI

[raullen @Raullen]

Is a flat $200/mo coding agent subscription actually worth it? Let's do the napkin math. 🧮

Agents are notoriously input-heavy (resending your repo every turn). A heavy month = ~250M input & 6M output tokens.

Priced out on DeepSeek V4.1 Flash:
• 250M Input @ $0.30/1M = $75
• 6M Output @ $1.20/1M = $7.2
• Total: ~$82/mo (peak) or ~$41 off-peak

You're paying 2.4–5x less for a model that hits 98% of GPT-6 Astra’s benchmarks according to @OpenDesignHQ

[Browser Use @browser_use]

Browser Use CLI makes your Hermes agent 10x better. Make sure to update to latest version!!

[Browser Use @browser_use]

https://t.co/pKTpXDUiqb

[Zachi @iam_zachi]

Huggingfaces security txt after the openai incident 😭