Jade Han

windowhan.eth | Lead Security Auditor @zellic_io prev. @kalos_security

[Jade Han @windowhan]

I want to do better in the new year. There are many areas I need to improve on. I also want to create more memories with the people I work with.

[Immunefi @immunefi]

There's a new legend in the making...

Researcher @windowhan snatched ELITE status last year in just under 10 total reports!

Congratulations ser... ⚔️

Will we see him and @deadrosesxyz go head to head in 2024?

[Kirill @kirillk_web3]

instead of watching 2 hours of Netflix tonight, watch this 40-minute masterclass from the founder of a $20B China AI company

it's the clearest explanation I've seen of how Agent Swarms and AI systems actually work at scale

useful whether you've never built an agent in your life or have been using Claude every day for the past year

I took the key ideas and turned them into a practical guide on how to actually build with Kimi

find it below

[Jade Han @windowhan]

@KKDW_KOREA @GONOGO_Korea 우와 이거 결말이 어떻게되엇나요 신박하네요

[Jade Han @windowhan]

@0x15_eth @Zcash @ipwning @ZcashCommGrants Congrats!

[Jade Han @windowhan]

@Krongggggg 동의합니다 추가로 더 대두되는 문제는... 그래서 진짜 취약점이 묻히기더하구요 ㅠ

[Jade Han @windowhan]

@gf_256 I'm always rooting for you. Just ignore the weirdos who are jealous of you. I hope your heart stays unhurt

[Robert Chen @NotDeGhost]

what a ridiculous take. @gf_256 / Zellic have done a ton for the space (speaking as one of their competitors).

did you even read the report?

[Taeyang Lee @5unkn0wn]

Everyone focuses on memory corruption bugs in the Linux kernel, but we shouldn’t overlook logical bugs.

[Jade Han @windowhan]

I saw an impressive piece of work someone created for the public good. I'd love to share it with others 😄

https://t.co/51HzLtD5Dw

[Jade Han @windowhan]

@developermin1 관심있습니다.

[Jade Han @windowhan]

@AnticsDecoded oh really? nice work 😄

[Jade Han @windowhan]

@devkoriel 익명으로 하심이...

[Jade Han @windowhan]

@devkoriel 공개적인 sns에 이렇게말씀하셔도되나요...ㄷㄷ 걱정되는..

[Jade Han @windowhan]

@AnticsDecoded @immunefi No! Only manual audit 🙂

[Jade Han @windowhan]

I just found a confirmed bug on @immunefi #immunefitribe https://t.co/RoAWJcCAbu

I achieved it after a very long debate/argument
And... yet another thing that got buried 🙃

[Jade Han @windowhan]

When doing bug hunting with Claude Code, I tried using various good skills and tools to improve my workflow, but the results were not as great as I expected.

After thinking about it, part of the issue was that I wasn't doing great prompt engineering or providing the right context — but also, since it's basically a black box in terms of what actually gets sent to and received from the LLM Provider, I had no idea what the problems were or which direction to improve things.

So I built a separate tool for debugging purposes, and I figured there might be others out there with the same need, so I'm sharing the repo: https://t.co/OT47Yc7wWg (vibe-coding 100%)

Features I'm thinking of adding later:
- Intercepting requests being sent to the LLM Provider so a human can intervene mid-process and adjust the direction
- Routing specific requests or specific sub-agent actions to a different LLM Provider
- Attaching a management Agent to monitor and supervise whether Claude Code has sufficient context while running, and check if anything has been missed

if anyone have additional ideas, feel free to me <3

[Jade Han @windowhan]

@whitenationkor 와 이런세상도잇네요

[Jade Han @windowhan]

I'm writing a whole drama here. I will win this war, no matter what. 👽

[Jade Han @windowhan]

@farazsth98 Let's goooooooooo

[Jade Han @windowhan]

@IFdLRx4At1WFm74 @bemusicscript zzzz...