A cluster of security disclosures on November 10 has drawn fresh attention to infrastructure vulnerabilities across the cryptocurrency and artificial-intelligence sectors, with lightning-network software, cross-chain bridging interfaces, and AI agent platforms all reporting active incidents within hours of one another.
LNbits, a bitcoin lightning-network payment software provider, released version 1.7.0 on November 10 as an urgent security update following active attacks that began on November 8, according to Bitcoin Magazine. The publication reported that the attacks drained funds from operators running prior versions of the software. In its advisory, LNbits instructed operators that they must upgrade to the patched release before restarting their servers, warning that earlier versions continued to put funds at risk. The expedited release reflected the severity of the active exploitation, which specifically targeted the lightning-network infrastructure layer used by merchants and node operators to process bitcoin transactions.
In the cross-chain bridging space, Across Protocol urged users to avoid its application amid widespread reports of drained wallets, as Cointelegraph reported. The protocol disclosed that it is investigating a potential security incident affecting its domain. Several users reported losing funds after interacting with the platform, the outlet noted, prompting an explicit warning to stay away from the app while the team assessed the scope of the compromise. The reports of drained wallets suggested that attackers may have gained control of frontend communications or routing, though the protocol had not yet confirmed the root cause at the time of the report.
On the artificial intelligence front, NEAR AI disclosed a breach tied to an SGLang vulnerability that exposed user details and data from 18 AI agents, Decrypt reported. The flaw allowed an attacker to reach internal systems and potentially view confidential AI requests, exposing the limits of hardware-backed privacy protections, the publication said. In a separate assessment, The Block reported that NEAR AI found no indication that stored Private Chat conversations or uploaded files were accessed during the incident. Nevertheless, the breach compromised user account details and raised questions about the resilience of privacy safeguards in AI infrastructure that handles sensitive queries at scale.
The convergence of these incidents within a single day underscores how infrastructure providers across Web3 and adjacent AI services remain high-value targets for attackers, with each case emphasizing the importance of rapid patch deployment, continuous domain security monitoring, and rigorous review of access controls around confidential user data.