escrow retweeted
We triggered WhatsApp 0-click on iOS/macOS/iPadOS.
CVE-2025-55177 arises from missing validation that the [Redacted] message originates from a linked device, enabling specially crafted DNG parsing that triggers CVE-2025-43300.
Analysis of Samsung CVE-2025-21043 is also ongoing.
escrow retweeted
Hello,
I've received a bunch of notifications today about the "Block Blaster" ... pseudo-takedown that occurred in response to a group of individuals spearphishing and cryptodraining a cancer patient.
I appreciate everyone thanking me or giving me a congratulations.
I am not fully responsible for the actions which occurred. I did reverse engineer the malware and identify infrastructure, however any work done was accelerated due to a group of people.
When I announced I was going to look at the video game closer to determine if it was malware (it was malware), a person contacted me and spun up a group of like minded people interested in examining Block Blaster closer.
Here are the cool and badass people I worked with:
- @zachxbt
- @John5725424446
- @andreee_eeeeee
- @escrow_
- @C4L38
- @downsin
- "J"
- Random nerds who provided "tips" to us
I've never really spoken with these people before, omit ZachXBT, but each of us was angry from what we had seen.
Before I get off for the evening I want to note that I am uploading Block Blaster to the malware library.
"./Samples/Families/Block Blaster"
I have also synced all samples in Triage and VirusTotal if you want to examine them closer. I noted the SHA256 hashes in a previous post.
escrow retweeted
The full technical report details how we took down the C2 infrastructure of Block Blasters drainer malware that @valvesoftware allowed on their platform. We spoke to the threat actors who had no remorse for stealing from a terminally ill cancer patient
docs.google.com/document/d/1…
escrow retweeted
tl;dr of today
> @rastalandTV gets crypto drained
> he has stage 4 cancer
> hes targeted specifically for his cancer treatment money
> loses $32,000
> nerds band together
> @ZssBecker donates $30,000 to him
> malware nerds come together
> drainer infra found
> pull all victim data from infra
> victims will be notified
> all malware flagged
> osint nerds come together
> find drainers info from their telegram ids
> find info from their steam ids
tl;dr tl;dr stage 4 cancer bro gets fucked over, 50+ nerds band together to undo the damage
fuck cancer
escrow retweeted
A detailed analysis of the Cork Protocol hack. @Corkprotocol
A DeFi "Insurance" protocol is not secure itself, turned open bar for attackers.
Cork Protocol — a platform meant to insure users against depegging.
Instead, it depegged from reality.
1. First, What is Cork Protocol?
An insurance protocol allows users to hedge the risk of depegging. It lets users bet for or against a stable asset holding its peg.
To achieve this, Cork Protocol has a set of markets where each of them composes 4 key assets and a PSM.
- RA: redemption asset. The price target asset.
- PA: pegged asset. The asset that is pegged to RA.
- DS: Depeg swap. The users who want to protect their token from depegging will buy DS. It's like an option to purchase RA with PA at a pegged rate.
- CT: Covered Token. The users who bet there is no depeg by selling the DS (similar to selling a call option)
The PSM has 2 functionalities:
- 1 DS + 1 PA => 1 RA (exercise of the option)
- 1 DS + 1 CT => 1 RA (redemption)
2. What did the Cork protocol hacker do?
0) There's a legit market A deployed by the project itself, where:
- RA = wstETH,
- PA = weETH,
- DS = weETH8DS-2,
- CT = weETH8CT-2.
1) creation of a fake market where RA=weETH8DS-2 and PA=wstETH. We denote the CT=ct_B and ds=ds_B.
DS tokens from the real market (weETH8DS-2) were now being treated as RA in this rogue market. A shell game begins.
2) The attacker legally buys weETH8CT-2 from Market A. Nothing suspicious. Yet.
3) Here’s the tragic twist.
Cork’s routerState contract at 0x55b9... was built to split incoming RA into corresponding DS + CT tokens for users. It assumes RA is a base asset, not a token from another market.
But the attacker fed it DS tokens (weETH8DS-2) from the real market — which, in the fake market, were considered RA.
RouterState blindly split them — as if they were fresh RA — into fake ds_B and ct_B.
For free.
Like watching a vending machine spit out sodas for tokens it shouldn’t even recognize. All this happened inside a Uniswap v4 hook.
4)
With stolen ds_B and ct_B in hand, the attacker converted them back into weETH8DS-2 via the fake market.
Wood -> splinters -> pass through a fence -> glue back into wood.
5) The Final Act
Now holding:
- weETH8CT-2 (from Step 2),
- weETH8DS-2 (reacquired via reassembly),
The attacker redeemed both through Market A’s PSM:
DS + CT => wstETH.
Game over.
What began as a depeg insurance protocol ended in a logic depeg.
---
Thanks to @RektHQ for using the similar tone from your articles.
---
I wanted to make this post more accessible after my earlier thread was too brief.
Let me know if I missed any technical details or I made any mistakes — happy to dive deeper.
#DeFi #hack #attack #cork
I cannot believe I am seeing a "judge" trying to disprove a researchers PoC on the premise of "artificially minting" the token that is required for the attack.
There are no restrictions to acquire such token in a real environment.
Do you realise how crazy this is?
> open @cantinaxyz to see if there is an update on my report
> there is a blue dot in notifs
> view the report
> severity nuked from crit to info
> no reason/justification provided from the judge
> ?
Don't miss out on @windhustler's offer.
Extremely skilled in any environment regarding LayerZero/Arbitrum integrations.
Highly recommend.
To demonstrate @burraSec's expertise, we’re offering a free full-day security review/consultation for projects integrating with LayerZero or Arbitrum—whether you’re already deployed or still in development.
We’ll thoroughly review:
LayerZero: Configuration (DVNs, Executor, and overall integration), functionality (LzRead, OFTs, vanilla OApps, and more).
Arbitrum: Native bridge or token bridge integrations, use of retryable tickets, or custom Orbit chains (e.g., custom gas tokens, USDC bridge standard).
DM me to schedule your review!
escrow retweeted
My notes for @windhustler interview on @deliriusz_eth channel
- "You are paid to think"
- Another example of switching fields, this time from engineering to IT, learning from courses
- 2018 java programmer
- 2021 web3 job, solidity developer
- 2023 web3 security
- Bug classes specialization
- Cross-chain apps
- layer zero
- 63/64 gas forwarding vulnerabilities
- Most interesting bugs
- Tx ordering (unconventional design in layer zero)
- Permits front-running to DoS the original tx
- Business-logic dependent
- If setters controlled by admins are called at unusual times, like in the middle of a user flow, what happens? Can it be front-run?
- Advice
- Drill through external integrations
- Explore parts others are not looking at
- Don't approach competitive audits like regular private audits, because uniqueness matters more than severity
- Focus on one part and try to find something unusual
- About contests
- Rules of contests
- Unclear, very subjective. A lot of people are annoyed by them
- Bugs often underinflated by judges, same as in web2
- Conditional rules have a problem: if you find A, make sure it does B, otherwise it's invalid
- Big contests' escalations never look too healthy
- Most people quit contests because they hate arguing with judges and navigating all the rules (move to private audits)
- But the escalations are very hard to solve, as well as judging as a service
- Only 3 severities (L, M, H) may not be enough
- Conditional pots give projects motivation to argue as much as possible
- Contests mostly rely on newcomers; without them, they wouldn't exist, as most participants move to private work after 6-12 months
- Judges often don't have enough time to argue and understand
- Always rushed
- But it onboarded a lot of people
- Opinions on platforms
- @windhustler
- Don't like payments for only the first, like @HatsFinance
- Leads to spamming
- Super rushed
- @cantinaxyz and sponsors seeing the issues during the contest
- It's fine because you don't have to submit. Leaves freedom to researchers
- Most experienced SRs submit in the last 30 minutes. Because something theoretically can happen (like malicious team member), no reason to take additional risks
- Less experienced have an opportunity to talk with the sponsor early when they are unsure
- But it's debatable if it's good when the first submission is rewarded with an additional bonus
- @escrow_
- Overall likes first come first served, (@HatsFinance)
- No dupes is good, bigger payments
- Faster payouts, less arguing
- But time constraints can get into your head and work against you, reducing mental clarity
- As soon as repo opens you see 40 spam submissions
- but other than that not much
- all these 40 instant submissions got invalidated
- @sherlockdefi live judging
- Fast payouts
- @deliriusz_eth
- Needs time to get a picture, find bugs, so doesn't like FCFS
- But it can be beneficial for those who have the skills to find bugs fast
- @windhustler ideal platform
- Lead auditor, like in @sherlockdefi. Most projects really like it
- Similar to reserved auditors on @code4rena
- Find really good judges
- Outsource judging
- Some QA, but not too crazy, like 50 messages for one issue
- @deliriusz_eth ideas
- Like in @code4rena, first validators invalidate 0-value issues. Helps judges a lot, reduces number of issues to check
- Add guaranteed auditor, like in @sherlockdefi
- Ranking system, where you can fall very fast if you do badly in some contests and lose LSW and the money for it
- Bigger pot for non-LSW participants
- About @immunefi boost
- Judging is better than on other platforms
- The most unique
- Requires PoC
- PoC is actually required in web2 too, most agree it should be required
- But some top auditors don't know how to write them or don't want to
- @deadrosesxyz do mostly manual reviews, very rarely PoC
- In some contests the winner finds 20-30 valid bugs, almost impossible to write PoCs for all of them
- Has duplicates
- SRs to look up to
- @panprog great auditor overall, especially for perpetual protocols
- @HollaWaldfee100
- @cmichelio
- @deadrosesxyz
- @zachobront
- immunefi.com/profile/GothicS…
- Not visible in the space, but has out of this world reports
- @0xriptide
- @trust__90
- AI
- Won't replace us, but will help audit, won't be possible to do audits without it
- Learning it will give a significant advantage
- Now can't find anything useful, but can explain some parts of the code, especially low-level
- Personal brand on X
- Some strategies
- Be really funny
- Review novel bug classes
- Genuinely share your struggles and experiences
- A lot of work, 1-2 hours a day cumulative, including time for comments on other posts
- Writing about something important to you helps clarify your mind
- Everyone should market themselves - if you're technically inclined, focus on writing in-depth technical content
- Typical day
- Weekday
- Wake up
- A few hours of work: talks, messages, tweets, audit
- Gym 1h
- Lunch 1h
- Another chunk of work, 4-5 hours
- Family time
- Some evening work, often just communication
- Weekends
- Some work, 3-4 hours
- Not every day is perfect, but tries to keep consistency
- Sometimes spend several hours organizing, planning and deciding what to do
- Constantly mix technical with non-technical work
- How to grow
- Read reports
- Work in teams
- Bug breakdowns in @SpearbitDAO
- But mostly just do audits and that's more than enough
- Approach to understand codebase
- Different every time
- Starts
- Sometimes starts from the smallest library and audits it in isolation, tries to understand only it first
- If already familiar with the protocol that the project integrates with, starts there. Easier to start with familiar and then expand
- User flows
- Tests help to figure them out faster, see the data
- Then build up from there
- Pattern-matching using known findings, e.g. from @soloditofficial, and then try to force this issue
- Use detectors that match the code with known vulnerable code
- @BowTiedDravee has some custom ones that show points of interest, like a static analyzer
- Checklists
- Private vs contests
- No need to argue about issues
- Assess the client, sometimes it's best to suggest redoing the architecture, not finding singular issues, because of all the complexities
- Usually team work
- Learn a lot from the team
- Discussions
- Very collaborative
- Health, including mental
- Gym
- Social life, family
- Find enjoyment every day
- Walks
- Less social media
- Remember that imposter syndrome is very common in our field, maybe more common than burnout
- Music during audits
- Classical can help, piano
- Some beats that you don't focus on, jazz
- But usually after 30-40 minutes you don't want to listen to it anymore, focus more on the audit
- Closing advice: focusing on specific protocols and platforms can help you advance quickly
- youtube.com/watch?v=xiIZyX0T…
escrow retweeted
In latest Threatside podcast we were joined with @windhustler , who shared his amazing road to success in Web3 🔥
watch it no YT youtube.com/watch?v=xiIZyX0T… 🍿
Or if you prefer other podcast platforms, check them out here: rss.com/podcasts/threatside/…
Thanks, as always, to my amazing co-host @escrow_ 🫡
escrow retweeted
Recently the bounty team at TrustSec found another critical leading to live unauthenticated theft of funds. Due to what we consider malicious behavior of the project and especially of @immunefi , not only did the project get away without paying the bounty, but due to a dirty loophole we are prohibited from publishing the details and sending bad public karma on the dishonest parties.
Effectively, by abusing the shady publication category 3 (immunefi.com/responsible-pub…), projects can get away with basically anything and honest hackers have zero leverage from public scrutiny. In fact, as you'll see we would have been better off without a bounty platform whose only action was protecting the project by securing our silence. Let's discuss the non-technical details:
- We identify a critical theft-of-funds and submit a PoC of it (no theoreticals, forked mainnet).
- Project confirms the issue but through nonsense argument declares it OOS and offers tiny good will bounty.
- We request mediation assuming any competent triager will side with us (this isn't a close call, with exp of judging 5000+ submissions on various platforms, we know it's an instant confirm). Publicly trusted figures like @zachobront can corroborate the severity.
- Shockingly we find out the Immunefi triager fell victim to project's rhetoric and declares it OOS and not eligible for awards.
- Project says offer of $X is still on the table.
- Immunefi urges us to accept it, we reject and state that by accepting it becomes a "Fixed and paid report" which doesn't allow us to publish the details without approval. We rather expose the scam and warn hackers than having a few extra Ks in our pocket.
- Immunefi says that it now treats the report status as "closed by Immunefi", which disallows publishing. That for us is the most sinister of all decisions in the entire message chain, because it's clear to all sides the issue is real, important (they offered a bounty for it) and assisted the project. But to save face, Immunefi is willing to be intellectually dishonest and call this "closed by Immunefi", as if it's a non-issue.
We're going public because the shady, ultra-secretive behavior we're seeing from projects and some bounty platforms goes directly against the Web3 ethos and the white hat community. Instead of celebrating openness and transparency, white hats are divided and conquered, fighting their own battles for justice against well-funded parties and without public leverage.
Immunefi, is anything written here wrong? Then publicly call out our BS by releasing the bounty transcript. Otherwise, your silence is your admission of guilt. Spare us any vague and empty statements which cannot be challenged without the technical details.
Retweet if you want something to change in this imbalance of power.
escrow retweeted
Web3 Builders:
As you know I’m bullish af on @RareSkills_io educational content but not all of you know they also have a sister organization called @RareTalent_xyz.
RareTalent screens applicants to ensure that only the highest quality engineers are selected within their resource pool and then made available for protocols development needs.
I’m going to be working with RareTalent to connect with and research protocols that are actively building and understand their needs from a staffing / strategy perspective.
Having trusted and highly skilled developers will reduce the probability of exploitation and will ensure effective operations of the protocol.
If anyone in my network wants to have a discussion around current or future staffing needs, shoot me a direct message.
escrow retweeted
For aspiring web3 security researchers and developers, @RareSkills_io educational content is a key resource that should be utilized in your learning journey.
While reading the chapters of the books, I thought it would be cool if you could have some accompanying CTFs to practically test the knowledge you have learned. So we built some!
Check out the repo below for the first CTF challenge based on Chapter 1 of the Uniswap V2 Book: ERC4626 Vaults.
Thanks @Jeyffre for providing feedback and helping us create this! Looking forward to creating many more challenges.
github.com/BlockChomper/ctf-…
escrow retweeted
had a really nice talk recently with @deliriusz_eth and @escrow_
highly suggest all beginners to listen to this as I shared a lot of my insights on how to be a better auditor 👀
New episode of Threatside landed 🔥 This time we speak with @deadrosesxyz , super star security researcher. Bugs literally tremble when they see him 😁 You can check it out here: rss.com/podcasts/threatside/…

























