The volunteer initiative Bitcoin Red Team, uniting 16 security specialists from around the world, has presented the first results of a large-scale audit of the first cryptocurrency's ecosystem. In the first 29.8 hours of work, the team recorded 4,962 potential issues across 390 bitcoin-related projects. Of these, 85 findings were classified as critical, and 635 as high-risk.

It is important to emphasize: only 21.4% of the discovered issues could be reproduced in a controlled environment. The remaining results cannot yet be considered confirmed vulnerabilities—they are merely hypotheses requiring verification. Nevertheless, the sheer speed and scale of the findings raise serious concerns.

What exactly the team is examining

The objects of analysis include cryptographic libraries, wallets, infrastructure software, and other open-source projects. The team includes well-known figures, such as Cashu protocol developer under the pseudonym Calle and AnchorWatch CEO Rob Hamilton. The methodology combines AI models with manual review: specialists wrote tools that automatically download code, identify dangerous sections, and simulate potential attacks in a local environment.

This is classic red teaming—imitating an attacker's actions. Researchers do not just look for errors but also assess whether they can be exploited to compromise a product. According to Calle, on average, each participant finds roughly one critical exploit per hour. He does not disclose the calculation methodology, which leaves questions about the accuracy of the statistics.

The list of affected projects is not published. Critical reports are first reproduced using a working attack example, then privately shared with developers. "There is a lot of chaos in the ecosystem right now. We understand that many are literally being flooded with security reports," Calle noted, commenting on coordination difficulties.

Cause for alarm: the Coldcard incident

The campaign launched against the backdrop of an exploited vulnerability in Coldcard hardware wallets. Based on my estimates from Galaxy Research analysis, at least 15 attackers took advantage of the flaw. Confirmed losses from three waves of attacks exceed $100 million, and with the suspected fourth wave, the total could reach $130 million.

Coinkite's technical analysis showed that the problem arose after a firmware change in 2021: in some cases, devices used a predictable MicroPython software generator instead of the hardware RNG. The effective entropy of seed phrases dropped from 128 bits to approximately 40 bits on Coldcard Mk2 and Mk3, and to 72 bits on Mk4, Mk5, and Q. Fixed firmware has already been released, but it does not protect already-created seed phrases—owners will have to generate new ones and transfer funds.

AI accelerates discovery, but not resolution

Hamilton emphasizes that the campaign's main limitation is not finding issues but delivering reports to the right developers. So far, the team has spent about $20,000 on various services, but funding is secured. Interestingly, this case echoes recent AI achievements in security: Anthropic's Claude model identified 22 vulnerabilities in Firefox within two weeks, 14 of which Mozilla recognized as high-severity.

My comment: This initiative is a wake-up call for the entire industry. The fact that dozens of critical issues can be found in a single day points to systemic weaknesses in many projects that have been considered reliable for years. In the short term, we can expect a wave of emergency fixes and, likely, new incidents. In the long term, this is an inevitable stage of ecosystem maturation, where security must become not an option but the foundation of development.