Sid

building Claude Code @AnthropicAI

previously @robinhoodapp @rubrikinc

[Sid @sidbid]

Super excited to finally share Dynamic Workflows in Claude Code!!

We built this a couple months ago, and it has slowly become a daily driver for a bunch of people at Anthropic. A few tips for getting the most out of it 🧵

https://t.co/WtwkSd3JPp

[ClaudeDevs @ClaudeDevs]

New in Claude Code (research preview): dynamic workflows.

Claude writes an orchestration script on the fly, then spins up a large fleet of coordinated subagents in parallel to take on your most complex tasks.

Use the word "workflow" in a prompt to get started.

[Sid @sidbid]

@navigateny It is expensive to run but you can ask claude to make your workflows smaller if you wanna test the waters

[Sid @sidbid]

Workflows have changed the way I use Claude Code. Once it clicks, it's hard to go back

[Thariq @trq212]

Workflows are the biggest upgrade to Claude Code’s capabilities since skills and subagents.

I dove deep into it with @sidbid to figure out best practices, examples and more.  I’m particularly excited about the non-technical tasks it enables for Claude Code.

[Sid @sidbid]

@PawelJLisowski can you /feedback and send me your id? i'd love to look into this for you

[Nityesh @nityeshaga]

I just turned one of our most token hungry skills into a dynamic workflow - now it uses upto 5x fewer output tokens while spawning more subagents!

This means significantly lower AI costs for this job while increasing the reliability of the workflow purely from changes in the harness level. If this was running this on API pricing, savings would be massive - while increasing the reliability!

Anthropic's insight of replacing an orchestrator from a non-deterministic AI agent to deterministic code written by an agent is such a simple and brilliant insight, it's blowing my mind! 🤯

So if you have tried to create skills to let your AI agent work do complex work, especially non-technical, it's time to ask your Claude to turn it into a Dynamic Workflow and see if that improves the outcomes.

[Sid @sidbid]

@jdjohnson You can disable this with /config today and we are going to change the trigger word soon as well

[Sid @sidbid]

@peteryxu Glad you're enjoying it!

[Sid @sidbid]

Highly recommend trying out Dynamic Workflows if you get a chance. I also can't imagine working without them anymore

(They are expensive to run, but you can ask claude to make them smaller to get a feel for token usage)

[dan @irl_danB]

I see a lot of people saying GPT-5.5 is still better than Opus-4.8

whether or not that's true

dynamic workflows has again changed my behavior so dramatically already that it doesn't matter. will be hard to return to codex until they have an equivalent

https://t.co/V5ecJPyRBr

[Sid @sidbid]

@dexhorthy @andrew_r @walden_yan @tobi @karpathy fun fact: the agents do have structured output and the model can theoretically route based on this :)

[Sid @sidbid]

@ignacioaal @jarredsumner It should know what "adversarial reviews" mean

[Sid @sidbid]

@mhlakhani @PawelJLisowski ♥️ You were one of the first true believers in the power of Workflows

[Sid @sidbid]

Super excited to finally share Dynamic Workflows in Claude Code!!

We built this a couple months ago, and it has slowly become a daily driver for a bunch of people at Anthropic. A few tips for getting the most out of it 🧵

https://t.co/WtwkSd3JPp

[ClaudeDevs @ClaudeDevs]

New in Claude Code (research preview): dynamic workflows.

Claude writes an orchestration script on the fly, then spins up a large fleet of coordinated subagents in parallel to take on your most complex tasks.

Use the word "workflow" in a prompt to get started.

[Sid @sidbid]

@GilFeig Run workflow --> /workflows --> select your workflow --> hit 's' --> enter name --> Your workflow has been saved and can be re-run with a slash command

[Sid @sidbid]

Normal code review:
Find bugs. Look harder and find more. 45 "bugs" later -- how do you know which are real?

Adversarial code review:
Find bugs. Keep going. For each bug, ask 2+ Claudes in isolated context windows to refute - is this bug real? Only the real ones survive.

Adversarial reviews improve accuracy, but cost more tokens. Ask CC workflows to run an adversarial review and it'll do the right thing!

[Sid @sidbid]

@maloymediika @ClaudeDevs Just type "workflow" in your prompt and it should light up

[Sid @sidbid]

@rhmoller @ClaudeDevs No this uses your existing subscription usage

[Sid @sidbid]

@alex_norcliffe You can switch off the keyword match per prompt today but I hear you. Will make this configurable

[Sid @sidbid]

@den_spirin Yeah if you hand it over to claude it can probable create a workflow out of it for you

[Sid @sidbid]

@trq212 ♥️

[Sid @sidbid]

@ali_tlekbai @bcherny Agents can communicate with each other, but that communication is encoded up front.

For example it might create a "bug finder" agent, and then a "verifier" agent -- and wire them up so the right information is passed along between them

[Sid @sidbid]

@PawelJLisowski Yes you can combine /goal and workflows (although be warned it may guzzle tokens)

If you can share examples of /goal giving up too soon I can have a look for you. Use /feedback and send me the ids

[Sid @sidbid]

@chufucious Try "launch a workflow to review my code from different perspectives"