Bitget App
Trade smarter
Buy cryptoMarketsTradeFuturesEarnSquareMore
web3: Bitcoin red team uses Chinese AI models to detect open-source vulnerabilities

web3: Bitcoin red team uses Chinese AI models to detect open-source vulnerabilities

币界网币界网2026/08/14 00:08
Show original

A volunteer team specializing in identifying Bitcoin software vulnerabilities claims to have conducted a basic scan of almost the entire Bitcoin open-source ecosystem using Chinese AI models, submitting numerous security issue reports to developers.

Calle, the head of the Bitcoin Red Team and a pseudonymous developer, stated on social media that the team combines AI tools with manual reviews to inspect projects such as wallets, Lightning Network applications, and software libraries. For verified issues, the team privately notifies developers to fix them before deciding whether to disclose details publicly.

Covered 390 Projects

Calle revealed that in August, the team submitted 4,962 issue records to 390 projects, including 85 severe vulnerabilities and 635 high-risk vulnerabilities, with the remainder being security issues of various levels.

He noted that developers have confirmed the existence of several genuine severe and high-risk vulnerabilities among these reports. However, the team has not yet disclosed the names of affected projects or any technical details.

Chinese Models Used for Code Auditing

The models used by this team include Kimi K3 from Chinese startup Moonshot AI and GLM 5.2 from the Chinese developer platform Z.ai, as well as models from OpenAI and Anthropic.

According to Calle, US-produced models have usage restrictions in security research scenarios, and the team frequently encountered limitations during the audit process, leading them to shift to Chinese models that can be run locally. Kimi K3 is described as capable of handling large codebases and executing lengthy software analysis tasks with minimal human intervention.

Lightning Network Project Audits Are More Challenging

Calle said that Lightning Network software, which supports fast Bitcoin payments, is notably more difficult to audit due to its complex structure, and it generally presents more significant issue patterns than typical projects.

He also mentioned that projects which adopted AI auditing processes early now differ markedly from those that have not implemented such tools. In the future, project teams will need to establish their own AI auditing processes, while software lacking regular maintenance should not be easily trusted.

Based on his statements, AI is boosting the efficiency of vulnerability discovery, but it is also increasing the pressure on developers to ensure security. Although he described the current situation as “Bitcoin is burning,” he believes that large-scale audits will ultimately make Bitcoin software more robust.

0
0

Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.

Understand the market, then trade.
Bitget offers one-stop trading for cryptocurrencies, stocks, and gold.
Trade now!

You may also like

Ivy League’s space gamble: Harvard reveals $2.2 billions SpaceX position, its largest single stock holding

Harvard University Investment Management Company disclosed that it holds $2.2 billion worth of SpaceX shares, making it the largest single stock position in its $4.3 billion US equity portfolio. This holding originated from Harvard’s early investment in SpaceX through venture capital funds, and the position has significantly appreciated after SpaceX’s public listing in June this year. The University of California also disclosed a roughly $1 billion position in SpaceX during the same period. Over the past decade, endowment funds from several universities have made substantial bets on private tech assets.

华尔街见闻2026/08/16 03:01