某地址因签署钓鱼签名损失超72万美元valBUSD与valTUSD

In March, there were 46 crypto security incidents, with losses totaling $39.8 million, setting a new high since November 2024

Gate News message, April 3, CertiK monitoring data shows that in March 2026 there were 46 security incidents recorded (excluding phishing), with total losses of approximately $39.8 million. This is the highest number of recorded security incidents in a month since November 2024. CertiK said that after hitting a low point in the third quarter of 2025, security incidents increased again in the fourth quarter of 2025 and the first quarter of 2026, with more and more instances of code vulnerabilities being exploited over the past two quarters.

SlowMist releases a security assessment report: A major Web3 wallet has found no risk of private keys or seed phrases being leaked

Blockchain security firm SlowMist released a security assessment report for a mainstream Web3 wallet. The results show the wallet does not transmit private keys or seed phrases to external parties. Overall, no sensitive data leakage risk was detected, providing users with a safe reference.

The Trump family’s crypto project was rated as D-level, with a high-risk warning for WLFI

Trump family crypto project WLFI was rated “D” by new rating agency CORE3, making it a high-risk investment. CORE3 provides public, transparent risk assessments by analyzing the likelihood of capital losses, pointing out that WLFI faces multiple risks, including a lack of monitoring and security measures. This rating has drawn investors’ attention and may affect market decisions.

Korean criminals use USDT to provide revenge services; Telegram becomes a new tool for anonymous crimes

South Korean police disclosed that some criminal groups use USDT to provide retaliatory attack services. Victims can pay in cryptocurrency to hire others to carry out retaliation, which has led to multiple cases. Police say it is difficult to track down the perpetrators. Experts warn that such services may spread as Web3 technology grows, and they call for increased cybersecurity awareness.

DeepMind warns: Six types of cyberattacks can hijack AI agents—companies need to strengthen protection

Google DeepMind researchers, in a report titled “The AI Agent Trap,” warn that an open internet environment could be used to manipulate autonomous AI agents. The report identifies six main risks, including content injection traps and behavior control traps, and recommends addressing them with measures such as adversarial training and behavioral monitoring, urging the industry to strengthen AI safety safeguards.

Pi Network forcibly enables 2FA to protect wallet security; migrating to the mainnet is no longer high-risk.

Pi Network rolls out a security update, requiring Pioneers to enable two-factor authentication (2FA) to protect users’ account and wallet security and prevent loss of funds during the mainnet migration. This move addresses ongoing scam threats and ensures that sensitive actions are approved only by the account owner. At the same time, node operators have upgraded to protocol 21.2 to improve system performance. Users must complete the mainnet migration step by step to safeguard their assets.