claude code was just leaked so obviously... i used multiple claude code agents to break down their code and find the secrets and there are plenty of them that im sure anthropic never wanted us to see

Using Claude to hack Claude is the most cybersecurity thing I've seen all week.

using claude code agents to reverse engineer claude code is the most beautifully recursive thing i've seen this week

There are posts that x402 is injected into the leaked code by someone else. And it winds up silently in places it shouldn't. If you are not running that agent in an isolated sandbox, you may want to double-check that. Especially as it looks like you're into trading.

did you clone it? i didnt. just gave the gh link

Ah, I thought on the video you're running a leaked version with "--dangerously-skip-permissions"

may i know what is the name of your IDE for this whole terminal management?

moon dev app, you get it when you come to the zoom

AI analyzing AI’s leaked code… we’ve officially entered recursion in real life.

What’s interesting isn’t that people are digging into it. That was inevitable What’s interesting is what they think is a “secret” vs what is just good system design

Yeah, they leaked it.

Using Claude Code agents to reverse-engineer Claude Code is beautifully ironic. I need to see what they uncovered.

Running multiple agents to reverse-engineer agent code is beautifully recursive. The real question is whether any of those findings survive the next patch — or if this is just a snapshot of a moving target. Meanwhile, we're out here using leaked tools to coordinate humanitarian infrastructure under blackouts. Different priorities, same species.

A few weeks ago, wasn't it said that you were going to disappear and retire, and that we had one last chance to pay you a super discount on your courses?

naw, its not discounted anymore sorry: retirement is overated id rather keep making yall cope

@thecryptoroad I wanted to attend your next $5 zoom session 😭😭💔💔

@thecryptoroad thats still $5

If you need to use their leaked code you have 0 creativity and none of their secrets will save you

Just watched this Claude code leak breakdown. I ran my own agents on it last night. The hidden background agents and multi-Claude orchestration are exactly what context management has been missing. Persistent memory across sessions solves the biggest fail I see in autonomous setups – everything resets when the run ends. From my engineering experience, study these leaks. They show the runtime that actually ships production code. No fluff. Build on it.

This is the public assassination of a company following recent US blacklisting Organized crime syndicate running our country demonstrated its power by decapitating an $800bn company Response and Escalation leading to civil war while 🍿

This is better

Using Claude agents to analyze Claude Code is the most meta thing this week. The multi-agent pattern is legit though — different agents examining different layers catches things a single pass misses

We now miss the “leaked…” sticker

这下可真是热闹了啊，Claude的秘密大公开？想知道代码里藏着啥秘密，看来得好好研究研究。别忘了分享发现，别全自己吞了啊～

could it be a massive april fools joke?

Where can I download the leaked source code

Using AI to dissect AI's own leak. But ask yourself: did you find secrets, or breadcrumbs they wanted found?

I don't care what they are hiding from me.

Or just have a look here?

did u see $buddy sir? claude just leaked its whole code and there is claude buddy which is unreleased. check this out! 👀👾 99RtLpiXfU57Xoedwnxg2D6UKjk8BsG3ES876z2kpump

Dm to get featured