Anthropic’s new advanced tool calling is gold and I’m surprised not many people talk about it

- Programmatic tool calling
- Dynamic filtering
- Tool use examples
- Tool search
…

Here is a quick 3 min break down 🧵👇 1/ Programmatic tool call

Most of us have been using LLM to glue together functions via tool call

Instead of forcing LLM output tool calls, just give it an environment that have access to all tools and write script to call those functions

Anthropic shows achieving ~37% token reduction

Similar research done in Code actions paper & cloud flare code mode

Vercel's agent-browser is wild

- 70% less token consumption compared with Chrome dev tool MCP in my test 🤯

- {bash + CLI} is SOOO much better than MCP

(Chrome MCP tools took 4k tokens VS agent-browser CLI instruction in CLAUDE.md is only 95 token)

More people will start building CLI instead of MCP, MCP is a dead concept

Below is more detailed break down comparison if you interested 👇 1/ Using Chrome dev tool MCP

- Tools itself took 4k tokens (2% context window)

- Tested prompt "help me go check whats the prompt of the first prompt in prompt library of " -> 38k token consumedsuperdesign.dev

Gemini 3's hidden power = Awesome UI Animator

It actually contains good knowledge around popular animation libraries

You just need right prompt spec + process to unlock it

Here is my process 🧵👇 There are 2 popular libraries:

1/ GSAP: Great for complex & scroll based animation
2/ Motion.dev: Great for React & UI state animation

However if you just prompt Gemini 3 like "Give me an awesome animation",

it will give you super boring & generic slide deck type animation

You need to prompt it specifically, describing the animation clearly, then it will surprise you 🤯

Nano Banana + Gemini 3 = S-Tier UI Designer

I've been using Nano Banana alongside Gemini 3 to maximise its creativity for UI design past 2 weeks

🧵👇 Here are my 4 step process 1/ Design spec

Plan ahead, Break down into smaller design tasks, and get model focusing one task at a time

This is a proven process we've been doing with coding tasks, same thing apply for UI design too

Here are a few sites I often go for reference...

Okay at this point,

I can only say Gemini3 really beats me as a designer...

🧵👇 Below is how I get most out of Gemini3 1/ Prompt for Gemini3 is different

Gemini 3 is a reasoning model

It requires precise instructions + Really sensitive to the prompt

Some old prompt engineering techniques that works with old model might not work well with gemini 3

Sometimes more prompt you add, the worse the performance

Claude Skill is prob bigger than MCP

Here is what it is,

+ how to 10x your claude code w/ it

🧵👇 1/ A skill ≈ prompt + tools/assets

.claude/
└ skills/
└ [skill-name]/
├ SKILL.md
└ (other sub-prompts, tools, templates)

SKILL.md firstly contain a simple yaml contains metadata,

Provides just enough info about when to use this skill

+ Full markdown prompt explain detailed instructions, which will only be fed into context if agent decide to use this skill

-----

For complex skills, in the full markdown instruction,

You can refer to more md files for further instructions

+ Scripts/functions you predefined OR any other resources

So that it can call predefined functions on-demand

.agent docs made my Claude Code 10x better…

Here is how to context engineer into Claude code

It even works w/ complex codebase

👇 Thread below 1/ Context engineering is the key

Claude code default takes in
- System prompt
- System tools
- MCP tools
-
- Messages

Run /context will give you an idea of the breakdown CLAUDE.md

Nano Banana is insanely good

Here are a few ways I found that make the best out of Nano bananas:

1. World knowledge

Because Gemini has world's knowledge, it goes beyond just image editing

You can prompt a point on map, and ask it to picture what the view will look like

"What will I see standing at this point on the map" from @tokumin
Similar to the world knowledge use case

You can also ask it to edit image with additional info

"You are a location-based AR experience generator. highlight [xxxxx] in this image and annotate relevant information about it."

from @bilawalsidhu

I was using Claude Code Sub-agent wrong...

Here's what I learnt after 20+ hrs trial & errors

Made my agents way more effective

👇 Thread below First -

Why did Anthropic build sub-agents initially?

It's all about context engineer & context optimisations

It works best when focusing on planning & researching and send back with a summary plan to conversation thread

I was using Claude Code wrong...

Here’s what I learnt and how I maximise Claude Code performance

+ Best tips that ACTUALLY useful

👇 Thread below The Claude Code setup
- Connect IDE
- /init for codebase scan
-
- ... CLAUDE.md

Referral/affiliate marketing is a hugely underutilized growth channel in web3

It generated:
- 30% of volume for Lido
- $11B trading volume for GMX

I analyzed 6 web3 projects to try to understand how BIG this could be.

Here's the data I collected 🧵 @LidoFinance referral program:

Generated 1,552,458 staked ETH from referral (~30% of total staked ETH)

Distributed $13M rewards distributed to affiliates past 18 months

dune.com/LidoAnalytical…