Read the latest Pulse

London edition

News checked through

Published 26 Oct, 07:20 UTC

You are reading a saved edition, not live news. News window: 25 Oct, 07:15 UTC to 26 Oct, 07:15 UTC. Labelled ongoing and upcoming items retain their own dates.

The catch-up

News from the last 24 hours, plus clearly dated ongoing developments and upcoming events.

Worth watching, unconfirmed

Bloom says nine users were hit after Solana wallet drains

Bloom is a trading bot for Solana and EVM chains. On 26 October @OnchainLens said address 7kY8qx8MqJ7Rz2vMhgHETdZLHcTp1oM78AA9dmE9p8BQ received 9,408 SOL, about $954,000, from 49 wallets, including one transfer of 1,636 SOL, about $166,000. It said the total might be higher. @BloomTrading then said those 49 wallets belong to nine users, not 49 separate people. It said the cause is unconfirmed, including whether a third party was involved, and that it will repay verified losses. It told users to move leftover funds to new wallets. This desk did not trace the address.

Read story

Robinhood

Checked, quiet

No material development selected for this edition's coverage window. This is not an exhaustive account. Checked 26 Oct, 07:15 UTC.

Solana

1 development

Bloom says nine users were hit after Solana wallet drains

Bloom is a trading bot for Solana and EVM chains. On 26 October @OnchainLens said address 7kY8qx8MqJ7Rz2vMhgHETdZLHcTp1oM78AA9dmE9p8BQ received 9,408 SOL, about $954,000, from 49 wallets, including one transfer of 1,636 SOL, about $166,000. It said the total might be higher. @BloomTrading then said those 49 wallets belong to nine users, not 49 separate people. It said the cause is unconfirmed, including whether a third party was involved, and that it will repay verified losses. It told users to move leftover funds to new wallets. This desk did not trace the address.

Why it matters A bot that can trade from a user’s wallet can also move those coins if that access is abused. Bloom’s count of nine users does not match the 49-wallet tally, and the project has not named a cause, so the size and source of the loss are still open.

Read the context

Bloom describes itself as cross-chain trading automation for Solana and EVM networks. Connecting a wallet to a bot often lets that bot send transactions, which is different from approving every trade by hand. @OnchainLens is an onchain-data account, not Bloom and not Solana, so its SOL and dollar figures are a reported tally rather than a Bloom audit. Bloom’s replies say the team has been active since 2021 and that verified losses will be repaid, but they give no payment date, no final SOL total and no technical cause. Solana’s own account did not post about this incident in the posts reviewed. Anyone still using the bot can compare Bloom’s replies with the address above.

Link to this story

Base

1 development

Base node notes name 5 November for faster blocks. Docs do not.

Base is Coinbase’s Ethereum layer 2. A new block there now arrives about every two seconds. On 22 October the base/base v1.5.1 release notes said operators must upgrade by 5 November 2026 at 18:00:00 UTC for the Denim hardfork, and that this release adds Denim on mainnet. Base’s Denim docs still say mainnet is only targeting November 2026, call that a planning target rather than an activation time, and say the required mainnet node software is not published yet. Those docs say 200 millisecond blocks are already live on the Sepolia test network from 21 October 2026 at 18:00 UTC, and on an experimental network called Vibenet. The 5 November time is what the release notes announce. It is not a guarantee, and the docs have not adopted it. This desk did not run a node.

Why it matters If the faster schedule turns on, a contract that treats a block count as a waiting period would finish much sooner. Operators have a dated upgrade note, while the public docs still say there is no activation time and that the software is unpublished.

Read the context

Base is an Ethereum layer 2 incubated by Coinbase. Denim would produce five complete blocks per second instead of one about every two seconds. block.number would advance about ten times faster in clock time. block.timestamp would still move in whole seconds, so a timer written with block.timestamp would not speed up. A timelock, vesting schedule, voting window or deadline written as a number of blocks would. Flashblocks, Base’s current early-confirmation stream, would stop once Denim is active. The docs also say the Sepolia and Vibenet versions can change how developers test, and that Vibenet should not be treated as production. The GitHub notes and the docs page were both still public at this check and they still disagree.

Link to this story

Arc

Checked, quiet

No material development selected for this edition's coverage window. This is not an exhaustive account. Checked 26 Oct, 07:15 UTC.

Elsewhere in crypto

1 development

Initia names a script-cache bug and says the chain is still halted

Initia is a blockchain built to host connected application chains, called rollups. On 24 October @initia gave a fuller account of the 23 October halt. It said that at 14:59 UTC an attacker used a bug in Initia MoveVM’s script caching to bypass security checks and credit unbacked WETH, iUSD and USDC without changing those tokens’ total supply. Native INIT was not minted. Before the halt at 15:41:30 UTC, the attacker bridged about $2.7 million: 105 WETH and about 1.4 million AUSD0 through LayerZero, and about 1.08 million USDC through Noble and Circle’s CCTP. Initia said the tokens were swapped for 1,135.52 ETH, now at Ethereum address 0x358ef0130f7a5aa516dcd81bdc80b18cde41cac8. The layer 1 and rollups were still halted. A patch and audit were expected in a few days, with no restart time. This desk did not use the network.

Why it matters The network and its rollups are still stopped. The project has now named the bug, the assets and an Ethereum address it says holds the proceeds. A restart is hoped for, not scheduled. This later project post is more specific than the earlier co-founder note.

Read the context

Initia’s first post on 23 October said only that the layer 1 and rollups were halted after a security incident and warned readers away from unofficial links. A co-founder then described a previously unknown MoveVM bug and estimated that about $2.7 million of AUSD, ether and USDC had been bridged out, while native INIT was outside the bug. The 24 October project post is the account to use: it names script caching, lists WETH, iUSD and USDC as the unbacked credits, and separates those credits from the bridged WETH, AUSD0 and USDC. It does not say users will be repaid, and it does not give a block height or a restart clock time. “A few days” for the fix was not a date. No later @initia post in this check showed that the chain had resumed.

Link to this story