Developers believe the vulnerability could lead to user fund losses. All nodes running versions 26.04 and above are at risk. As an immediate measure, the project team recommends suspending routing and payment processing. For those unable to fully shut down their node, it is advised to restart the background process with the “offline” parameter. This mode disables connectivity with other Lightning Network nodes—payments cannot be sent, received, or routed—but the software will continue running and monitoring the Bitcoin blockchain.

The vulnerability has been assigned a CVE identifier, but experts have not yet clarified the technical details of the flaw. A two-week embargo on reporting the issue is intended to give operators time to apply patches before attackers can analyze the changes.

This is not the first time AI has detected flaws in the Bitcoin network. At the end of July, a group called Bitcoin Red Team, comprising 16 developers, used AI models to analyze 390 Bitcoin project repositories and announced that they had identified nearly 5,000 security issues, 85 of which were classified as critical and 635 as high-severity vulnerabilities. However, no detailed descriptions of the vulnerabilities were provided.
The Lightning Network operates on top of the Bitcoin blockchain, enabling faster transactions with lower fees by not recording every payment on the main chain. To do this, two participants lock a certain amount of bitcoins into a shared payment channel and repeatedly update the distribution of funds between them. When the channel is closed, the final balance is recorded on the blockchain. A node must continuously monitor the network: if the other party attempts to close the channel using outdated data, the node can respond on the blockchain to secure its funds. A completely powered-off machine cannot do this.
Previously, Charles Guillemet, Chief Technology Officer of hardware wallet manufacturer Ledger, stated that a vulnerability affecting clear signing in the Ethereum app had been patched. The flaw was discovered by the Ledger Donjon team using an AI-powered vulnerability detection system.

