An AI-powered security scanner developed by Jack Dorsey’s Block identified a vulnerability that could temporarily prevent Bitcoin nodes from processing network messages. The issue involved repeatedly sending the same transaction ID up to 50,000 times, potentially causing a node to become unresponsive for several seconds.
Bitcoin Core, the software responsible for operating these nodes, disclosed the vulnerability on Friday under the identifier CVE-2026-117348. Developers also released versions 30.5 and 31.3, which include security patches to address the issue.
According to an investigation by CoinDesk, the security advisory acknowledged an automated AI-powered code scan conducted through Project Loupe, led by Elias Rohrer at Spiral, Block’s open-source development division.
Bitcoin nodes are computers that validate transactions and relay payments and newly mined blocks across the network.
The vulnerability could have affected several parts of the cryptocurrency ecosystem. For example, exchanges operating vulnerable nodes might experience delays when processing customer deposits. Similarly, miners could waste computing resources by working with an outdated version of the blockchain.
Bitcoin Core developers classified the vulnerability as having medium severity and being easily discoverable. Under the security policy introduced on October 22, vulnerabilities that are easy to identify receive prompt fixes across supported software versions, accompanied by public security disclosures.
How the Bitcoin Node Vulnerability Worked
Bitcoin nodes temporarily store orphan transactions, which are transactions that attempt to spend coins from an earlier transaction that the receiving node has not yet encountered. The software searches for the missing parent transactions to complete the verification process.
Since Bitcoin Core version 29.0, the software has handled announcements of orphan transactions by creating a list of missing parent transactions and checking its existing records before requesting the information from other nodes.
However, Bitcoin nodes can announce transactions in batches containing as many as 50,000 entries, and the software did not automatically remove duplicate transaction IDs from these announcements.
An attacker could exploit this weakness by initially sending an orphan transaction with thousands of missing parent transactions. They could then broadcast an announcement containing the same transaction ID repeated 50,000 times.
This would force the receiving node to perform the same resource-intensive checks repeatedly rather than processing the transaction only once.
According to the security advisory, each malicious announcement could freeze a node for more than a dozen seconds. During that period, the affected node would stop handling messages from other network participants, including announcements of newly mined blocks. Repeated attacks could extend the disruption.
Bitcoin Core’s latest patches address the problem by ensuring that each transaction included in an announcement batch is processed only once, regardless of how many times its ID appears.
The incident also highlights growing concerns among Bitcoin developers about the security implications of AI-powered vulnerability discovery. As automated tools make it easier to identify software weaknesses, attackers could potentially rediscover publicly documented or previously patched flaws more quickly. This has increased the importance of releasing security fixes across supported versions without unnecessary delays.

More Stories
XRP Drops 10% as Long Liquidations Mount and ETF Assets Fall Below $1B
Bitcoin Crashes Below $57,000 for First Time in Two Years as Crypto Sell-Off Deepens
Bitcoin Extends Losses for Third Day as Asian Tech Stocks Retreat