Technology•

Bitcoin Core Patches a Node-Freezing Flaw That Block's AI Scanner Loupe Uncovered

Bitcoin Core released versions 30.5 and 31.3 to fix a medium-severity bug that let one repeated transaction ID freeze a node for over a dozen seconds, found by Block's AI scanner.

Written By :MINRK
Bitcoin Core Patches a Node-Freezing Flaw That Block's AI Scanner Loupe Uncovered

Photo by Kanchanara on Unsplash

Mentioned in This Article

BTC
BTC
Bitcoin
$59,027.00
2.12%

1. A Bitcoin Core vulnerability found by an AI scanner

Bitcoin Core published a security advisory on November 20, 2026, and shipped versions 30.5 and 31.3 to fix a Bitcoin Core vulnerability, tracked as CVE-2026-117348, that let an attacker stall a node by repeating a single transaction ID 50,000 times in one message. The flaw was found by Loupe, an AI-driven code scanner from Block's open-source unit Spiral.

The developers behind the most widely used Bitcoin software rated the problem medium severity. It did not put coins at risk or allow anyone to rewrite the blockchain. What it did allow was cheaper and more mundane, but still disruptive: a remote peer could make a node stop processing network traffic for a stretch of time, and could keep doing so by sending more of the same messages.

The discovery also stands out for how it was made. Rather than a human researcher combing through source code, the issue surfaced during an automated scan run by AI agents, a sign of how quickly machine-assisted auditing is moving into the security work around Bitcoin's base software.

2. Who found the bug, and how was it credited?

The bug was credited to an automated AI scan carried out through Project Loupe, an effort led by developer Elias Rohrer at Spiral. Spiral is the open-source development arm of Block, the payments company co-founded and led by Jack Dorsey, and it funds and builds free software for the Bitcoin ecosystem.

Loupe is itself open source. It deploys AI agents to examine Bitcoin code and flag behavior that could be abused, the kind of painstaking review that has historically depended on a small pool of specialists with deep familiarity with the codebase. The Bitcoin Core advisory names the Loupe scan as the source of the report.

Block has long positioned itself as one of the more Bitcoin-focused large companies in payments, and Spiral is the part of that strategy aimed at the protocol's public infrastructure rather than at Block's own products. A finding like this one, routed through the project's disclosure process and fixed in an official release, is the type of outcome such funding is meant to produce.

3. What does a Bitcoin node do?

A Bitcoin node is a computer running software, most often Bitcoin Core, that independently checks every transaction and block against the network's rules. Nodes also pass transactions and newly mined blocks along to the peers they are connected to, which is how information spreads across the network without any central server.

Because nodes are constantly exchanging messages with strangers on the internet, the software has to be careful about how much work any single message can force it to do. If one peer can make a node spend a long time on a cheap-to-send request, that peer has a lever for slowing the node down. Security engineers call this a denial-of-service problem, and it is one of the main categories of risk that node developers guard against.

Nodes are not only run by hobbyists. Exchanges, custodians, payment processors and mining operations all rely on their own nodes to learn about incoming payments and new blocks. That is why a flaw that merely pauses a node, rather than crashing it or corrupting data, can still matter in practice.

4. Why a frozen node matters to exchanges and miners

While a node is stuck, it is effectively deaf to the network. It is not reading new transactions, and it is not reading new blocks. For a business, that gap can translate into real friction.

An exchange whose node was targeted could be slow to notice and credit customer deposits, since it would not see the relevant transactions or confirmations until the node resumed work. A miner relying on an affected node could end up building on an outdated picture of the blockchain, spending electricity and hardware time on a version of the chain that the rest of the network has already moved past.

Neither scenario involves theft. Both, however, impose costs, and an attacker able to repeat the trick could stretch those costs out. Delayed deposits erode customer trust, and wasted mining effort is money that never comes back. That combination of low cost to the attacker and real cost to the target is what places a bug like this in the medium-severity band.

5. How the Bitcoin Core vulnerability worked

The weakness sat in the way nodes handle so-called orphan transactions. An orphan is a transaction that spends coins created by an earlier transaction, its parent, that the node has not yet seen. Rather than throwing the orphan away, the node keeps it temporarily and tries to track down the missing parents so it can verify the full chain of spending.

Orphans are a normal part of how a decentralized network operates. Transactions travel from peer to peer along different paths and at different speeds, so a node can easily hear about a child transaction before the parent it depends on has reached it. Handling that situation gracefully, instead of rejecting valid activity, is part of what keeps transaction relay smooth across thousands of independently run machines.

Starting with version 29.0, Bitcoin Core changed how it reacts when a peer announces an orphan. The node works out which parent transactions are missing and checks its own records before asking anyone for them. That check is sensible on its own, but it is not free, and the cost grows with the number of missing parents involved.

Peers announce transactions to one another in batches, and a single batch can hold up to 50,000 entries. The flaw was that the software did not filter out duplicates in such a batch. If the same transaction ID appeared many times, the node would treat each appearance as a fresh item and repeat the full set of checks every time.

An attacker could exploit that in two steps. First, send the victim an orphan transaction with thousands of missing parents, making each lookup expensive. Second, send an announcement that repeats that orphan's ID 50,000 times. The node would then grind through the same costly work tens of thousands of times over.

6. More than a dozen seconds per message

The Bitcoin Core advisory spelled out the effect: each malicious message could freeze a node for more than a dozen seconds. During that window, the node handled no other messages at all, including new blocks arriving from the network.

Crucially, the pause was not a one-off. An attacker could keep sending further messages of the same kind to extend the freeze, turning a brief stall into a sustained disruption for as long as they cared to continue. Because the attack relied on ordinary peer-to-peer messages, it did not require special access to the target, only a connection to it.

This is the core of why denial-of-service bugs are taken seriously even when they cannot steal funds. The asymmetry is stark: one crafted announcement costs the sender very little, while it costs the receiver many seconds of complete unresponsiveness.

7. The fix in Bitcoin Core 30.5 and 31.3

The remedy is simple in concept. In the patched releases, a node processes each transaction in an announced batch only once, no matter how many times its ID is repeated. Duplicate entries no longer trigger duplicate work, which removes the multiplier the attack depended on.

Bitcoin Core delivered the change in two maintenance releases, 30.5 and 31.3, so operators on either of the currently maintained release lines can pick up the fix without jumping to a different major version. For node operators, and especially for exchanges, custodians and miners whose operations depend on timely block and transaction data, moving to one of those versions closes off the issue.

The patch illustrates a broader pattern in node security. Many denial-of-service flaws do not stem from exotic cryptographic failures but from ordinary assumptions, such as expecting a list to contain distinct items, that an adversary can deliberately violate.

8. Medium severity under Bitcoin Core's new disclosure policy

Core developers described the bug as both medium severity and easily discoverable. That second label carries weight under the security policy the project adopted on October 22, 2026.

Under that policy, flaws in this category are fixed quickly across all maintained versions, and a public advisory is released at the same time as the fix rather than months later. The approach is a departure from the more cautious habit of shipping a fix quietly and waiting before explaining what it addressed.

The reasoning ties directly to tools like Loupe. If AI systems can find a bug cheaply, then attackers using similar tools can rediscover it cheaply too. In that environment, keeping a fix quiet offers less protection than it once did, because the window during which only defenders know about a weakness is likely to be short. Publishing promptly gives operators the information they need to upgrade without pretending the secret will hold.

9. What AI-assisted auditing means for Bitcoin security

The episode offers an early look at how automated, AI-driven review could reshape the maintenance of critical open-source software. Scanners that never tire can sweep through code paths that human reviewers might only examine occasionally, and they can flag problems that are, in the developers' own words, easy to find once someone looks.

That cuts both ways. The same capability that helped Spiral's Loupe surface this flaw is, in principle, available to anyone with bad intentions. Bitcoin Core's decision to treat easily discoverable bugs as candidates for fast, public fixes reflects an acknowledgment of that reality.

For now, the practical takeaway is straightforward. CVE-2026-117348 has been identified, explained and patched, and node operators running affected software can address it by upgrading to Bitcoin Core 30.5 or 31.3.

Related Articles

NEWSLETTERS

Don't miss another story.

Subscribe to the MINRK Newsletter today.

Tick the newsletters you want. No selection? You'll get MINRK Headlines.

By signing up, you will receive emails about MINRK products and you agree to our terms of use and privacy policy.

MINRK Daybook

Market analysis for crypto traders and investors.

EVERY WEEKDAY

MINRK for Advisors

Defining crypto, digital assets and the future of finance for financial advisors.

EVERY THURSDAY

MINRK Protocol

Exploring the tech behind crypto one block at a time.

WEEKLY

MINRK Long & Short

A must read for institutions. Insights, news and analysis delivered weekly.

EVERY WEDNESDAY

MINRK Headlines

The biggest crypto news and ideas of the day.

EVERY WEEKDAY

MINRK Policy

Examining the intersection of cryptocurrency and government.

WEEKLY

MINRK Research

Join thousands of readers who rely on MINRK for data-driven insights on the latest digital asset trends.

MONTHLY