BTC $64,635.55, ETH $1,911.35, SOL $75.32, XRP $1.10, HYPE $58.81, DOGE $0.072605, AAVE $96.31.
Monday’s useful signal is specificity.
July 25 was pressure leaving the screen. July 26 was controlled access, then the cost of operating the door.
This morning’s cleaner read is more granular.
The market is no longer only asking who can enter the room. It is asking what each contract says, which deadline actually matters, which bridge key can drain funds, which open model is cheap enough to use, and which developer workflow can catch mistakes before they ship.
The CFTC warned prediction-market venues against broad template certifications. CLARITY still has roughly two weeks before the August recess turns policy into calendar risk. Ethereum ETF demand cooled without killing the rotation story. Bitcoin and Ethereum-linked protocols lost more than $35M across several attacks within hours. Thinking Machines Lab’s Inkling release shows open-weight AI is becoming an infrastructure fight, not just a benchmark fight. GitHub’s fresh tape is full of code review, coding surfaces, and Hyperliquid monitoring.
That is a different Monday mix. Less “who owns access?” More “can the details survive contact with regulators, attackers, users, and developers?”
Price snapshot via Coinbase spot prices for BTC and ETH, plus CoinGecko simple-price data for SOL, XRP, HYPE, DOGE, and AAVE around 04:20 HKT.
1. The CFTC Wants Prediction Markets To Stop Filing In Bulk
CoinDesk reported that the CFTC warned prediction-market firms against submitting broad, template-style event-contract certifications.
That sounds procedural. It is more important than that.
Prediction markets are trying to scale like software: build a category, reuse a template, add new markets quickly, and let distribution do the work. Regulators are pushing the other way. They want each contract’s settlement source, wording, manipulation risk, public-interest implications, and legal basis to stand on its own.
This is the next phase after Robinhood, Kalshi, Crypto.com, ForecastEx, and others turned event contracts into consumer-app supply.
The fight is no longer only whether prediction markets are legal. It is whether a platform can industrialize contract creation without turning regulated review into a rubber stamp.
If the CFTC holds this line, the winning venues won’t just have the best app distribution. They’ll have the best contract factory.
2. CLARITY Has A Calendar Problem, Not Just A Vote Problem
BeInCrypto reported that Senate leaders doubt the CLARITY Act will pass before the August recess.
That fits the weekend’s weaker odds.
July 25 focused on Galaxy cutting 2026 passage odds to 30%. The Monday question is sharper: what actually gets done before senators leave town?
Crypto bills don’t fail only because everyone votes no. They fail because negotiators run out of floor time, amendments pile up, ethics language becomes politically toxic, and leadership decides there is an easier fight elsewhere.
This matters for markets because delay preserves uncertainty. Exchanges, issuers, DeFi teams, custodians, and banks can price a strict rulebook. They struggle more with a rulebook that always looks two weeks away.
The number to watch is not 30%. It is the recess clock.
3. Ether Flow Cooled, But The Rotation Story Is Still Alive
The Block reported last week that ether ETFs had beaten bitcoin funds on net inflows for a second straight week while bitcoin ETF volume fell to its weakest full-week total since October 2024.
That does not erase last week’s signal.
Bitcoin ETF volume just hit its weakest full week since October 2024. Ether funds still beat bitcoin funds on net inflows last week, even from a much smaller asset base. This morning ETH is near $1,911 while BTC sits around $64,636.
The useful read is rotation fatigue, not rotation failure.
When a narrative starts working, traders immediately ask if it is crowded. A broken streak gives the market that test. If ether keeps holding relative strength while flows wobble, the rotation thesis becomes stronger. If it only works on clean inflow days, it is just a shorter trade.
The next data point is whether ETH can keep taking attention when ETF flow is no longer one-way.
4. Clustered Protocol Attacks Put Key Management Back In Focus
CoinDesk reported that Bitcoin and Ethereum-linked protocols lost more than $35M in multiple attacks within hours.
The clustering matters.
Crypto security usually gets narrated as one exploit, one postmortem, one patch. Attackers don’t operate that neatly. They reuse reconnaissance, hunt similar configurations, watch hot-wallet behavior, and move quickly when a weakness proves live.
This is why bridges, wrapped assets, admin keys, upgrade permissions, monitoring, and incident response matter as much as smart-contract audits.
A protocol can have clean public code and still lose money if operational keys, relayers, signers, or treasury workflows are brittle. The attack surface is no longer only “the contract.” It is the whole path money travels through before users think the transaction is final.
Security teams should treat clustered events as a warning siren, not isolated bad luck.
5. Inkling Shows Open-Weight AI Is Becoming Infrastructure
Decrypt reviewed Inkling, the first open-source AI release from Mira Murati’s Thinking Machines Lab.
The crypto angle is developer advantage.
Open-weight models change who can build private tooling, local assistants, research systems, compliance helpers, and security workflows without routing everything through a hosted provider. The winning stack won’t be the model alone. It will be the model plus memory, evals, permissions, tool routing, and clean product taste.
That is why the release matters even if traders don’t care about another AI leaderboard.
Open models let teams own more of the workflow. Hosted models still win when reliability, distribution, and product polish matter more than control. The real market is the hybrid middle: local where privacy and cost matter, cloud where quality and speed matter.
Crypto teams should pay attention because their data is sensitive, their workflows are adversarial, and their users get punished when automation is confidently wrong.
6. Treasury Rotation Still Needs Real Operating Discipline
NewToTheStreet’s syndicated release on Newsworthy kept public-company crypto treasury strategies in focus, noting Forward Industries expanded its Solana treasury by more than 500,000 SOL during fiscal Q3.
The market has learned to understand one bitcoin-treasury model. Everything after that is messier.
Solana, ether, and mixed-asset treasury strategies can produce cleaner upside in a rotation tape, but they are harder to explain. Investors need to know which asset is held, how it is financed, who gets paid first, what dilution looks like, and how management handles drawdowns.
This is the same lesson Strategy forced into the open with net reserve metrics.
A crypto treasury is not only a wallet balance. It is a capital structure wrapped around volatile collateral. The asset can be right and the security can still be priced wrong.
That distinction will matter more if alt treasury vehicles keep copying the bitcoin playbook without the same liquidity depth.
7. GitHub Trending - Fresh Picks After The Repeat Filter
The repeat tracker ruled out July 26’s anthropics/claude-cookbooks, palmier-io/palmier-pro, and OtterMind/Chat2DB, plus recent repeats like block/buzz, ComposioHQ/awesome-claude-skills, citrolabs/ego-lite, CoreBunch/Instatic, Pumpkin-MC/Pumpkin, shiyu-coder/Kronos, and yorukot/superfile.
alibaba/open-code-review has about 13.7K stars. It is a hybrid code-review system that combines deterministic checks with LLM review, including line-level comments and rules for null pointers, thread safety, XSS, and SQL injection. The signal is that code review is moving toward layered checks: rules catch known failure modes, models catch context, and humans review the judgment.
pingdotgg/t3code has about 15K stars. The repository description is sparse, but the trend is still useful: AI coding surfaces are becoming product categories in their own right. Developers don’t only want a chat box. They want repo context, edits, review, previews, and deployment feedback in one working surface.
0xhype/hyperliquid-tracker has about 511 stars. It sends alerts for large Hyperliquid trades. That is small compared with the AI repos, but more relevant to crypto users. As Hyperliquid absorbs more perp and RWA flow, the edge moves from “can I access the venue?” to “can I see meaningful flow before everyone else reacts?”
Morning Read
Read CoinDesk’s CFTC warning, then read BeInCrypto’s CLARITY calendar piece, then read CoinDesk’s clustered protocol-attack report.
The number to remember is $35M.
That is the reported loss across multiple protocol attacks within hours. The second number is two weeks, because U.S. market-structure legislation is now fighting the Senate calendar as much as it is fighting policy disagreement.
Monday’s read is specificity. Prediction markets need contract-level discipline. Crypto legislation needs calendar discipline. ETF rotation needs more than a clean inflow streak. Protocol security needs operational key discipline. AI tooling needs a real deployment stack, not just a new model. GitHub’s fresh repos point the same way: code review, coding surfaces, and trading alerts are all becoming more useful when they turn broad automation into specific, inspectable decisions.
Crypto’s next serious edge is not another vague platform claim. It is precise rules, precise risk controls, precise alerts, and precise tools that users can trust when the market gets noisy.
Evening Update - 18:44 HKT
BTC $65,186.54, ETH $1,962.79, SOL $76.40, XRP $1.11, HYPE $60.04, DOGE $0.072666, AAVE $101.05.
The evening read moved from specificity to operating claims.
Morning asked whether the details can survive regulators, attackers, users, and developers. Evening added a more concrete test: can the system keep working when macro risk cools, receivables move onchain, bank payments use blockchain rails, cloud-storage equity has to be rewritten in court, hot wallets keep receiving deposits after a drain, bridges learn that “the contract was fine” isn’t enough, and shared AI chats leak crypto-sensitive context?
This is a different mix from the last few digests. Less prediction-market supply, ETF rotation, venue exits, and sanctions plumbing. More business continuity, payment operations, bankruptcy design, offchain security, data leakage, and builder tooling that tries to make AI output less generic.
Price snapshot via Coinbase spot prices for BTC and ETH, plus CoinGecko simple-price data for SOL, XRP, HYPE, DOGE, and AAVE around 18:44 HKT.
8. Bitcoin Got A Cleaner Macro Tape, But The Week Is Still Loaded
CoinDesk reported that bitcoin moved back above $65,000 as the U.S. and Iran held fire and oil fell about 5%.
That is the cleanest market update of the day.
Lower oil reduces the immediate inflation scare. Ether led the majors, gaining more than bitcoin and pushing the ETH/BTC ratio to a three-month high in CoinDesk’s live market updates.
The catch is the calendar.
This week has the Federal Reserve, PCE inflation, second-quarter GDP, big tech earnings, Coinbase earnings, and Strategy earnings. A ceasefire bid can reopen risk appetite, but it doesn’t remove the liquidity tests stacked behind it.
The market’s best version is simple: oil cools, yields stop pressing higher, ETH keeps outperforming, and BTC holds the mid-$60Ks as the defensive anchor. The worse version is also simple: the Fed or AI-capex guidance pulls yields higher again, and crypto’s calm turns into another crowded relief trade.
9. POSCO Put Real Receivables Onchain
CoinDesk reported that POSCO International and LG CNS are testing trade-receivable tokenization on Injective using real commercial invoices.
This is a better tokenization story than another fund wrapper.
POSCO is South Korea’s largest trading company, with $22.2B of revenue last year across steel, energy, and battery-materials businesses. The pilot uses receivables from actual trade between overseas operations and counterparties, then tries to make one shared, transferable record for buyers, sellers, and banks.
Trade finance is full of boring friction. Everyone has a version of the invoice. Everyone reconciles at a different speed. Cash gets released only after records line up.
Putting receivables onchain doesn’t make credit risk disappear. It can make the claim easier to verify, transfer, settle, and finance. That is where tokenization earns its keep: working capital, not vibes.
10. KB Kookmin Is Bringing JPMorgan’s Blockchain Rails To Corporate Payments
The Block reported that South Korea’s KB Kookmin Bank plans to launch blockchain-based cross-border corporate payments next month using Kinexys by J.P. Morgan.
The first lane matters.
Yonhap said the service will be available through KB Kookmin’s domestic branches and Singapore branch, initially focusing on U.S. dollar transfers across 10 countries: South Korea, the U.S., Singapore, Saudi Arabia, India, Thailand, Qatar, the UAE, Bahrain, and South Africa.
That is not a retail crypto app. It is bank-to-corporate payment plumbing.
The point is predictability. Corporates care about settlement time, reconciliation, compliance, FX handling, and whether the bank can explain the product to auditors. If Kinexys becomes part of normal import and export settlement, blockchain stops being an innovation lab label and becomes a treasury operations feature.
South Korea now has POSCO testing receivables and KB Kookmin testing corporate payments in the same news cycle. That is a real corporate-finance cluster.
11. Storj Bankruptcy Tests Whether Token Holders Get More Than A Ticker
Cointelegraph reported via TradingView that Storj Labs filed for Chapter 11 protection while saying its decentralized storage network will keep operating.
The unusual part is the proposed tokenholder path.
Storj said it is exploring a court-approved ownership mechanism for STORJ holders. That turns an old crypto question into a bankruptcy question: when a token sits beside a company, infrastructure network, operating business, and user base, what do holders actually own when the company restructures?
The network can keep serving files while the corporate shell goes through court. That separation is useful, but it also exposes the ambiguity.
Token holders often get upside language during growth. Bankruptcy asks for a legal claim, not a community slogan. If Storj can give holders a court-approved equity route, it may create a template. If it can’t, the lesson will be harsher: token exposure and company recovery are still two different things.
12. Triple-A’s Hot-Wallet Drain Became A Deposit-Routing Failure
The Block reported that losses tied to crypto payments firm Triple-A’s hot wallets climbed to about $11.8M as new deposits kept landing in affected addresses and being swept.
The fresh consequence is not only the larger number.
It is the fact that deposits kept arriving.
Once a hot wallet is compromised, the incident response isn’t just “find the attacker.” It is freeze routing, warn counterparties, rotate addresses, stop automated deposits, update checkout flows, and make sure integrators aren’t still sending funds into a live drain.
Triple-A said it is investigating and that customer funds weren’t affected, according to The Block. The harder operational question is how fast a payment company can turn off every pathway feeding a bad address.
Crypto payments firms sell finality. In a wallet compromise, they need reversibility at the operational layer before the chain finalizes someone else’s mistake.
13. Garden Finance Shows Offchain Solvers Are Part Of Bridge Security
Cointelegraph reported via TradingView that Garden Finance temporarily disabled its app after Blockaid reported about $450,000 in USDT drained across Ethereum, Base, Arbitrum, and BNB Smart Chain.
Garden’s explanation is the useful part.
The team said the protocol and HTLC smart contracts were not compromised. Instead, an independent solver’s offchain database was breached, fraudulent transaction records were inserted, and solver-owned funds were released for swaps that had not been funded by the counterparty.
That is exactly the kind of failure bridges have to model.
Users experience the product as one bridge. The implementation is a stack: contracts, solvers, databases, relayers, monitoring, security vendors, app front ends, and human escalation. If one solver’s offchain state can leak value, then offchain state is part of the security boundary.
“No user funds lost” is good. It is not the same as “the bridge design is finished.”
14. Crypto Vaults Are Starting To Look Like A Regulatory Category
The Block’s Funding column asked whether crypto vaults are funds and whether curators are fund managers.
That is a good question because vaults have moved from DeFi-native UX into something regulators can recognize.
A vault takes deposits, applies a strategy, has parameters, may have a curator or risk manager, and can look to users like someone else is managing assets for them. The assets may be stablecoins, bitcoin, ether, or other tokens. The wrapper may be smart-contract code. None of that automatically answers the regulatory question.
This is where DeFi’s language starts to matter less than user reliance.
If a user is mainly trusting code they can inspect and control, the argument looks one way. If a user is mainly trusting a named curator, advertised strategy, fee model, and portfolio choices, it looks another way.
Vault builders should pay attention before enforcement turns a product category into case law.
15. Strategy’s Buy Pause Is Turning Into An Earnings Test
BeInCrypto reported that Strategy has skipped bitcoin purchases for four straight weekly filings, its longest pause in two years, and reports second-quarter earnings Thursday, July 30.
That is the clean follow-up to the morning treasury-discipline story.
Strategy’s model works best when its stock trades at a rich premium to its bitcoin holdings. That premium lets the company sell equity or preferred securities, buy more BTC, and increase bitcoin per share. When mNAV falls toward 1, issuing new common equity can become value-destructive.
BeInCrypto said Strategy sold 3,588 BTC in July to fund preferred dividends and cash reserves.
That flips the usual Saylor question. The market doesn’t only need to know whether Strategy still likes bitcoin. It needs to know whether the capital stack still lets the company buy without hurting common holders.
Thursday’s earnings call is not just about BTC holdings. It is about whether the machine still compounds.
16. Shared AI Chats Became A Crypto Privacy Risk
BeInCrypto reported that publicly shared Claude chats were discoverable through Google, with some users finding crypto wallet details, credentials, CVs, and company files in shared conversations.
This is not a model-risk story. It is a sharing-risk story.
Claude chats are private by default, but a share link creates a public page for the conversation up to that point. BeInCrypto said Google showed the links but not page previews, so the damage came when people clicked through to public conversations.
For crypto users, the asymmetry is brutal.
A password can be rotated. A seed phrase, private key, signing policy, wallet map, or treasury workflow cannot be made private again after it leaks. Even if no confirmed theft came from these indexed chats, the lesson is immediate: never paste wallet secrets, exchange keys, recovery plans, or private transaction ops into a chat you might later share.
AI tools are becoming part of crypto operations. Their share buttons need the same paranoia as wallet approvals.
17. GitHub Trending - Fresh Picks After The Repeat Filter
The repeat tracker ruled out this morning’s alibaba/open-code-review, pingdotgg/t3code, and 0xhype/hyperliquid-tracker, plus recent repeats including citrolabs/ego-lite, block/buzz, CoreBunch/Instatic, OtterMind/Chat2DB, shiyu-coder/Kronos, anthropics/claude-cookbooks, Pumpkin-MC/Pumpkin, and yorukot/superfile.
pbakaus/impeccable has about 51K stars and gained more than 400 today. It describes itself as a design language that makes AI harnesses better at design. The signal fits the week: AI coding is no longer only about producing code. Teams now need taste, layout rules, review heuristics, and product judgment baked into the harness.
andrewyng/aisuite has about 15.5K stars. It gives developers one simple interface across multiple generative AI providers. That is boring in the right way. As model choice changes weekly, serious teams don’t want every workflow welded to one provider SDK.
permissionlesstech/bitchat-android has about 6.9K stars and gained more than 250 today. The Android fork matters because Bitchat’s repo-level fight with India turned distribution into the story. When a censorship-resistant messaging tool becomes politically sensitive, ports, mirrors, clients, and install surfaces become part of resilience.
Evening Read
Read CoinDesk’s POSCO receivables report, then read The Block’s KB Kookmin payments piece, then read The Block’s Triple-A wallet-drain update.
The number to remember is $11.8M.
That is the reported Triple-A hot-wallet loss after fresh deposits kept feeding compromised addresses. The second number is 10, because KB Kookmin’s first Kinexys-based corporate payment route targets dollar transfers across 10 countries.
Monday evening’s read is operating claims. If tokenization is real, it should make receivables easier to finance. If bank blockchain rails are real, they should make corporate payments easier to reconcile. If decentralized storage is real, the network should keep running while the company restructures. If crypto payments are serious, compromised wallets should stop receiving deposits fast. If bridge security is serious, offchain solvers count. If AI tools are becoming part of crypto work, shared chats cannot leak wallet context.
The market got a friendlier macro tape today. The harder work is still operational: route the payment, protect the wallet, secure the solver, explain the vault, survive the bankruptcy, and make the tools less sloppy before users trust them with money.