vulnerabilities

Anthropic announced that the Claude Mythos Preview model has made breakthroughs in cryptographic research, discovering improved attack methods against the post-quantum digital signature candidate HAWK. HAWK is a post-quantum signature candidate solicited by NIST to combat quantum computer attacks, which has already passed two rounds of expert review. However, Claude reduced the effective key strength of HAWK-256 from 2^64 to 2^38 in just 60 hours, significantly lowering the theoretical cost of cracking. HAWK has not yet been deployed in practice, and this attack currently does not affect any production systems.Claude also discovered an improved attack against 7-round AES, achieving a speed increase of 200 to 800 times. During the research process, Claude independently completed literature reviews, mathematical reasoning, and experimental validation with limited guidance from cryptographers. The HAWK attack took about 60 hours and had an API cost of approximately $100,000; the AES attack was completed independently by Claude, which generated about 1 billion tokens before proposing core innovative ideas. Anthropic researchers then spent hundreds of hours validating the results. Anthropic stated that AI models can now help identify significant flaws in cryptographic algorithms, and the cryptography community may face bottlenecks similar to those in the software vulnerability field—AI-generated research results far exceed human verification capabilities. Anthropic has collaborated with academic institutions to launch the CryptanalysisBench benchmark and coordinated disclosures with NIST authors and government partners. The research team has achieved preliminary results on algorithms such as LEA and Serpent-128. Anthropic warns that as AI capabilities improve, actual attacks against deployed systems may be discovered in the future, necessitating the establishment of response mechanisms in advance.

According to Bitcoin News, the latest Senate draft of the CLARITY Act includes a provision that encourages ethical hackers, specifically "white hat" hackers, to responsibly disclose cybersecurity vulnerabilities. The proposal would authorize rewards for individuals who identify and report security flaws, aiming to enhance the protection of digital asset infrastructure before malicious actors exploit these vulnerabilities. This provision incorporates the views of former CFTC Chairman J. Christopher Giancarlo, who has long supported innovation in digital assets and market resilience.

The founder of Slow Fog, Yu Xian, retweeted a tweet on the X platform regarding the potential poisoning attack risks of Claude Code and published an analysis of the poisoning attack details targeting Grok Build CLI and Claude Code CLI.It pointed out that the security mechanisms of Grok Build CLI are not unified, with different code paths having different trust assumptions, creating gaps that serve as channels for attackers. Attackers may execute arbitrary commands through malicious project configuration files without the user's knowledge, thereby stealing API keys, cloud credentials, or controlling local devices.Researchers constructed a testing environment and found that on Mac systems, if Claude Code is affected, executing specific test commands can trigger the local calculator to launch, proving the existence of potential command execution risks.If the attack is successful, attackers may further steal API keys from AI services like Claude and OpenAI, resulting in account cost losses, gain access to servers and data by obtaining cloud service credentials from AWS, Alibaba Cloud, Tencent Cloud, modify code repositories to implant backdoors, and use local devices as jump points to attack corporate internal networks. It is reported that the related vulnerabilities have existed for a year.

According to CoinDesk, the Ethereum Foundation recently disclosed that its security team used AI agents to test the software running on Ethereum validator nodes and successfully discovered a remotely triggerable vulnerability that could cause node crashes. However, researchers emphasized that human review remains a key step in distinguishing real vulnerabilities from false positives among the large number of security reports generated by AI.It is reported that the discovered vulnerability exists in the Ethereum network message propagation protocol gossipsub, where an attacker can remotely trigger the node software to enter an abnormal computation state, causing the program to crash and shut down, making the validator node go offline until the operator manually restarts it. This vulnerability has been fixed and registered under the identifier "CVE-2026-34219."Nikos Baxevanis, a member of the Ethereum Foundation's protocol security team, stated that what was truly surprising about this incident was not the AI's ability to discover vulnerabilities, but rather the amount of time the team spent discerning which vulnerabilities were real and which were merely seemingly credible "illusions."

The Ethereum Foundation Protocol Security team researchers stated in a blog post on Thursday that a series of AI agents have been deployed to test the software relied upon by Ethereum, searching for vulnerabilities in the cryptographic systems, protocol code, and smart contracts. The vulnerabilities discovered by the AI agents include a remotely triggerable panic issue in the peer-to-peer layer libp2p gossipsub used by the Ethereum consensus client, which has been fixed and disclosed on GitHub as CVE-2026-34219.The researchers noted that the AI agents are organized into specialized roles such as reconnaissance, searching, patching, and verification, aimed at identifying potential attack paths, reproducing failures, and verifying their applicability to production code. The Ethereum Foundation stated that AI has not replaced security researchers but has changed the way they work, allowing the team to cover a much broader scope than manual reviews, although researchers need to exercise more careful judgment over a large number of seemingly credible conclusions.

The Zcash development team announced plans to introduce formal verification through the upcoming new privacy pool Ironwood, using mathematical proofs to eliminate the risk of "undetectable inflation vulnerabilities."According to previous news, although the vulnerabilities in the Orchard shielded pool have been fixed and there is no evidence of exploitation, the community decided to launch a brand new Ironwood shielded pool and conduct comprehensive formal verification of its protocol, as it is theoretically impossible to confirm whether hidden inflation has occurred through on-chain history.

According to CoinDesk, researchers from the blockchain security company Hexens have discovered a "stale cache" type obfuscation vulnerability in the Aptos blockchain Move virtual machine. Attackers only need about $3,000 in server costs to launch an attack in a simulated environment with nearly a 90% success rate, without requiring validator permissions or insider knowledge.In simulated tests, researchers ran about 20 attacks, successfully executing 17-18 times, and verified the potential control over cross-chain protocol management permissions such as LayerZero, Wormhole, and USDC CCTP.Hexens assesses that this vulnerability directly threatens DeFi, stablecoins, and liquid staking protocols on the Aptos chain, involving low single-digit billion-dollar assets; if spread through cross-chain bridges, stablecoin minting, and centralized exchanges, the systemic risk exposure could reach up to $70 billion. The Aptos team completed the fix and deployed it to the mainnet within hours of receiving the vulnerability report on February 25, and currently, no user funds have been compromised.

The Chief Information Security Officer of Slow Fog, 23pds, posted on the X platform that curl recently fixed 18 security vulnerabilities, involving issues such as authentication bypass, memory safety, and host verification, among which one libcurl vulnerability has existed for about 25 years. The risks are not limited to the curl command line but also widely affect applications, SDKs, containers, firmware, gateways, and CI/CD environments that rely on libcurl. It is recommended to upgrade curl/libcurl as soon as possible and check for the use of old versions of libcurl, especially paying attention to mTLS, proxy authentication, and connection reuse scenarios.

Web3 security company GoPlus Security pointed out that there has been a recent surge in smart contract vulnerability attacks, particularly targeting older contracts that have been deployed for many years. Attackers are leveraging AI technology to quickly discover and exploit these vulnerabilities.Recent typical incidents include:On June 9, a Token of Power (TOP) contract deployed 7 years ago was attacked on Ethereum, resulting in a loss of approximately $1.5 million;On May 25, a WUSD.fi contract deployed 3 years ago was attacked, resulting in a loss of approximately $200,000;On June 14 and June 18, Aztec Network was attacked twice in a row due to an old contract deployed 2 years ago, with total losses exceeding $4 million.GoPlus Security believes that traditional auditing methods are insufficient to effectively cover legacy contracts, and adopting AI-based Always-on Audit services may be the most effective solution currently available, capable of conducting security checks on old contracts in a short period while balancing reliability and cost.

According to the Associated Press, Anthropic's Mythos model has discovered vulnerabilities in the U.S. government's classified systems.

In response to rumors regarding the attack on the BNB Chain OLPC/LABUBU liquidity pool, PancakeSwap's official announcement stated that the platform has noted the news related to this security incident. After preliminary verification, PancakeSwap's own smart contracts do not have security vulnerabilities. The project team is still continuously investigating the complete causes of the incident and will promptly share the latest investigation progress once more clues are obtained. Meanwhile, the official reminds users that all relevant event information should be based on content released through PancakeSwap's official channels.

Zcash founder Zooko Wilcox posted to express gratitude to Anthropic for helping to protect Zcash users.At the request of Shielded Labs, they conducted a security audit of Zcash using Mythos, and no other serious vulnerabilities were found in the Zcash protocol. Shielded Labs and other teams are continuing to advance security reinforcement work, and updates will be released later.

The founder of Slow Fog, Yu Xian, stated that the attack Asterix encountered is similar to that of Flooring Protocol and BMP yesterday (the underlying protocols are DN404 and BT404), with high-level NFT ID displacement operations overflowing and being reused. It seems the attackers are looking for common vulnerabilities.It is reported that Asterix disclosed an attack incident affecting the ASTX token contract yesterday, stating that its Uniswap v4 liquidity pool was attacked on June 8, with attackers stealing approximately 30 ETH through 242 transactions. The vulnerability originated from a lack of token ID restriction checks on approval operations in an early version of DN404. Attackers exploited the outdated token approval to repeatedly sell tokens in the pool to obtain ETH, and then extracted equivalent tokens through forged IDs, leading to a cycle that drained the funds.Smart contracts are immutable and cannot be patched. The team advises users to stop interacting with the current pool and tokens and is planning to migrate to deploy secure tokens. The team suspects that the attackers used a jailbroken AI tool for fuzz testing to discover unconventional logic paths.

Dragonfly partner Haseeb tweeted to clarify the recently fixed vulnerability in Zcash. He stated that if the vulnerability had been exploited before the fix (although the probability is extremely low), it would manifest as the privacy pool minting false ZEC. Attackers would need to sell quickly, but since the market primarily trades transparent ZEC, they cannot directly sell the unmasked privacy ZEC on mainstream exchanges. Therefore, the losses would mainly be borne by users holding privacy ZEC, with almost no impact on holders of transparent ZEC and price discovery.Haseeb emphasized that the Zcash team will introduce a new turnstile mechanism and a brand new privacy pool in the next upgrade to verify that the privacy pool has not been inflated. He also has high hopes for formal verification technology, believing it is an important path for enhancing software security across the industry in the future, especially crucial for privacy protocols.

According to the THORChain blog, ZEC has been added to the THORChain launch queue, but due to a recently disclosed vulnerability in Zcash that affects the normal use of integrators, THORChain needs to complete a code modification in the Bifrost module before proceeding.The development team stated that the modification is very minor but must be completed before the launch of ZEC. Currently, Monero (XMR) is expected to launch at the end of this month, with ZEC following afterward.

The ZEC treasury company Cypherpunk posted on platform X in response to the market fluctuations of the ZEC token, stating that all software has vulnerabilities. Historically, Bitcoin has "over-minted" 18.4 billion BTC due to errors. However, this does not mean that blockchain technology should be abandoned; rather, security should be enhanced through formal verification and provable correctness.Cypherpunk emphasized that with the development of AI technology, vulnerability detection will become faster and broader, but the key lies in who can identify problems before malicious actors do. Zcash demonstrates its capability in this regard through an upcoming update.Previously, it was reported that the privacy coin ZEC faced market sell-offs due to the revelation of a security vulnerability that could potentially lead to unlimited issuance, causing the coin's price to drop over 50% in a single day.

On May 29, 2026, Taylor Hornby discovered a critical forgery vulnerability in the Zcash Orchard funding pool. Taylor Hornby reported the vulnerability to the Zcash Open Development Lab, and all parties collaborated to complete the fix on June 2. This vulnerability could be exploited to secretly create an unlimited number of forged ZEC within Zcash Orchard. Due to the privacy features of Orchard, it is impossible to cryptographically prove whether the vulnerability was exploited before the fix was deployed. The vulnerability had existed since the activation of Orchard in May 2022, until an emergency fix was deployed on June 1, 2026.With the assistance of AI tools, Taylor Hornby wrote a complete exploit and generated unlimited and undetectable forged ZEC in a local testing environment. Currently, Shielded Labs is collaborating with other Zcash developers to explore network upgrade proposals that would allow anyone to verify the integrity of Zcash's supply.

According to The Block, the Donjon security team under Ledger has bypassed the firmware verification mechanism of the TROPIC01 chip used in the Trezor Safe 7 through precise laser attacks in a laboratory environment, allowing unauthorized firmware to be loaded as long as the attacker possesses the physical device.Chip manufacturer Tropic Square further discovered that there are additional attack vectors in the MAC-and-Destroy security mechanism used for PIN verification on this chip, but details will not be disclosed until the enhanced version of the chip is launched at the end of 2026.Trezor stated that the PIN, mnemonic backup, and private keys have never been stored on a single chip, and they have notified partners, so ordinary users do not need to take action. Currently, the feasibility of attacks can be reduced by disabling the chip's MAINTENANCE mode.