GetChain News
中简 中繁 EN
GetChain News
Toggle sidebar

Online/Update

News linked to both this project and an event.

NGINX HTTP/2 Vulnerability Allows Remote Code Execution, Patched Version Released

Rapid7 disclosed an NGINX HTTP/2 vulnerability that allows unauthenticated attackers to execute remote code under certain conditions. An official fix has been released.

SlowMist CISO: NGINX Exposes Two Vulnerabilities That Could Allow Attackers to Execute Remote Code and Gain Root Access

SlowMist Chief Information Security Officer (CISO) 23pds has warned of two critical security vulnerabilities in NGINX servers (CVE-2026-123007 and CVE-2026-123633), which could lead to remote code execution and root privilege takeover risks.Among them, the most severe, CVE-2026-123007, allows unauthenticated attackers to execute malicious commands on affected servers when specific conditions are met. Both vulnerabilities reside in NGINX's HTTP/2 code and can only be triggered when HTTP/2 is enabled, which is not enabled by default in NGINX. NGINX is a widely used web front-end server that powers more than 225 million websites worldwide.

PyPI Quarantines Two Versions of Microsoft AI Framework Containing Data-Stealing Code

Security researchers discovered that two versions of Microsoft's agent-framework contain credential-stealing code, and PyPI has quarantined them.

Anthropic Adjusts Claude Code Prompts to Restrict Cloning Existing Games

Anthropic updated the system prompt for Claude Code on Opus 5.5, restricting the recreation of specific games by others, while original games and those using self-designed assets and materials can still be created.

Binance Wallet Upgrades Agentic Wallet Security Mechanism: Restricts QR Code Login on High-Risk Websites

Odaily Report: Binance Wallet has announced the completion of a new round of Agentic Wallet security upgrades. For security reasons, the platform now restricts users from logging in to third-party high-risk websites via QR codes based on real-time risk control rules; for specific websites that trigger risk control rules, users will be forcibly logged out. The platform will also conduct regular security scans.Binance Wallet stated that some users have recently reported QR code login failures. Binance Wallet reminds users that to protect asset security, do not scan QR codes from unofficial platforms to log in to Agentic Wallet, and only use trusted Agents such as Binance AI Pro and Claude, Codex, Grok, etc.

Anthropic Updates Claude Usage Policy, Prohibiting Unauthorized Model Distillation and Malicious Cyber Attacks

According to Anthropic's official website, the company has released a new Usage Policy that will take effect on November 12, 2026, applying to Claude, Claude Code, the API, and third-party integration services. The policy explicitly prohibits the unauthorized use of model outputs to train other AI models, including model scraping and model distillation, while also banning the use of Claude for malicious cyber attacks, malware development, and fraudulent activities.

Cursor: Security incident discovered involving an extension, estimated to affect up to 5,000 users

Odaily reports: Cursor, the AI coding tool under SpaceX, posted on X: "We have discovered a security incident affecting a tampered Claude Code VS Code extension accessed through our Cursor IDE extension marketplace, affecting up to 5,000 users. This malicious extension version was not developed by Anthropic, and it does not affect other extensions downloaded from the official website, Grok Bot, Grok, or our other SpaceXSI products.In response, we shut down the extension marketplace this afternoon, worked with the relevant service providers to remove the malware, and reopened the marketplace after confirming there was no further risk. If you are affected, please refer to the instructions in the official blog post to learn how to identify the malicious package and take measures to mitigate any potential impact. We have asked all customers to update Claude Code for VS Code to version 2.1.314."

PHP Officially Urges Users to Update Immediately to Avoid Severe Remote Code Execution Vulnerabilities

Odaily reports: PHP has issued a security advisory regarding two previously disclosed critical remote code execution (RCE) vulnerabilities, identified as CVE-2026-109618 and CVE-2026-109619. Attackers may exploit these vulnerabilities to execute malicious code without authentication, and PHP officially strongly recommends that all users upgrade immediately.These vulnerabilities affect PHP environments that handle web requests, including PHP-FPM, Apache mod_php, php-cgi, FrankenPHP, and LiteSpeed. Among these, Apache mod_php poses a particularly prominent risk, as attacks can be triggered before application code runs.Official patched versions have been released: PHP 8.5.12, 8.4.27, 8.3.36, and 8.2.35. Users still running PHP 8.1 or earlier should contact their distribution or long-term support (LTS) vendor to confirm patch availability, or upgrade to a supported version. Officials emphasize that temporary mitigation measures cannot fully eliminate the risk—installing security updates immediately is the only complete way to address the vulnerabilities.

CoreWeave launches Physical AI Launchables to simplify physical AI workloads

Odaily reports: AI cloud service provider CoreWeave has announced the launch of Physical AI Launchables, unveiled at the NVIDIA GTC Berlin event. The product offers physical AI workload templates built and tested by CoreWeave engineers, aimed at robotics, autonomous vehicle, and simulation teams, enabling task launch via a single command and reducing upfront preparation work such as environment configuration, storage setup, and task scheduling.The first batch of templates covers five tools—NVIDIA Isaac Sim, CARLA, MuJoCo, NVIDIA AlpaSim, and NVIDIA Cosmos 3—supporting workflows such as data generation, model evaluation, and policy training, and integrating Cosmos Data Factory-related agent skills usable for video, image, and text data augmentation, auto-labeling, and data curation. The platform also provides specially optimized skills for coding agents such as Claude Code and Codex, supporting automatic task launch, run-status monitoring, and result retrieval.CoreWeave is also collaborating with physical AI software vendors including Cadence, Encord, and Voxel51 to bring related products onto its cloud infrastructure.

Oracle Launches Industry Application AI Agent Studio, Supporting Codeless and Low-Code Development of Industry Agents

According to an Oracle announcement, the company has launched the Industry Application AI Agent Studio, a no-code and low-code platform designed for industry applications that can be used to create, test, orchestrate, deploy, and manage AI Agents and multi-agent workflows. Built on Oracle AI Agent Studio for Fusion Applications, the platform can access existing application data, tools, and business processes, and supports features such as natural language and visual orchestration, cross-application workflows, Model Context Protocol (MCP), multiple generative AI models, testing and debugging, security controls, lifecycle management, and monitoring.

SlowMist Security Alert: PHP Exposed with Two Critical Remote Code Execution Vulnerabilities, Official Patch Expected Tonight

SlowMist has issued a security alert stating that PHP released a security advisory on October 20, disclosing two critical security vulnerabilities that could lead to remote code execution (RCE), tracked as CVE-2026-109618 and CVE-2026-109619, both with a CVSS v4.0 score of 9.2. Currently, PHP officials have not yet released a fix patch, and the specific affected versions and exploitation conditions remain to be confirmed.According to available information, PHP officials plan to release a security update on October 21 at 14:00 UTC (22:00 Beijing time). SlowMist recommends that relevant organizations promptly check the PHP versions in their servers and business systems, with particular attention to PHP services exposed to the public internet, and upgrade in a timely manner once the patch is released; systems that cannot be upgraded immediately should restrict access sources and take isolation measures when necessary.

Anthropic Opens Office in Singapore, Signs MOU with Government

According to the Anthropic official website, Anthropic announced the opening of its fifth Asia-Pacific office in Singapore and signed a Memorandum of Understanding (MoU) with Singapore's Ministry of Digital Development and Information. The two parties will collaborate in areas such as AI safety, model evaluation, and responsible AI adoption. Singapore is one of Claude's most active global markets, where over a thousand government agencies and enterprises, including GovTech Singapore, DBS, Grab, and GIC, have already deployed Claude. Among them, over 1,500 engineers at GovTech Singapore have adopted Claude Code to advance government system modernization, while Grab leverages Claude-driven sales assistants to save approximately 40,000 labor hours per quarter.

Telegram 13.0 Version: Plans to Launch Cryptocurrency Wallet Gram Wallet, Original Wallet Service Renamed to Walt

According to Durov's Code, Telegram is expected to launch a new cryptocurrency wallet, Gram Wallet, in the latest version 13.0 update, though the service has not yet been officially opened to users. Meanwhile, the original Wallet service has been renamed Walt and moved from the settings menu to the mini-app interface, with existing accounts and assets remaining unaffected.

Security agency: Hackers used LLM combined with AI penetration tools to attack South Korean financial institutions, leaking data of approximately 68,000 people

security firm CrowdStrike released an intelligence report stating that an unidentified hacker, between late September and early October, combined large language models (LLMs) with the domestic open-source AI penetration testing tool ARTEX to launch targeted cyberattacks against multiple South Korean financial institutions and steal data.After analyzing the hacker's exposed open directory, it was found that they employed a dual-server architecture centered on a Hong Kong IP, and used models including Claude Code, GLM-5.3, Grok 4.6, and DeepSeek v4.1-flash accessed through an API proxy to orchestrate the attack process. The hacker also asked Claude how to monetize the leaked South Korean data in Telegram groups. According to estimates by South Korean media, at least seven South Korean financial institutions were attacked, including KB Kookmin Bank, Shinhan Bank, and Hana Bank, with personal data of approximately 68,000 people leaked, involving sensitive information such as annual income and loan limits.

Zcash Developers Bring Tachyon Code into Zakura Common, Aiming for Over 50,000 Private Payments Per Second

Odaily News: Zcash developers are integrating parts of the scaling project Tachyon's code into the Zakura Common cryptography library, with a new component called Udon released on Monday. Tachyon lead Sean Bowe stated that Udon was previously developed as part of Tachyon and has now been added to Zakura Common to leverage the library's existing performance optimizations.The long-term goal set by Bowe and Dev Ojha, lead at Valar Group, is to achieve over 50,000 private payments per second. Tachyon aims to reduce the amount of data that Zcash nodes need to receive and verify by merging proofs from multiple transactions into a single proof. However, the upgrade has not yet been activated on the Zcash network.According to ZecStats data, in the 7 days ending September 28, Zcash averaged approximately 6,200 transactions per day involving shielded funds, down from an average of approximately 11,700 per day in the 7 days ending September 21. (CoinDesk)

Solana Founder: New Project Percolator's Code Is Still Incomplete; Development Work Can't Be Fully Handed Over to AI at This Stage

Odaily reports: Solana founder Toly posted on X, stating: "AGI (Artificial General Intelligence) is still a long way off. As for my test project Percolator, if I were writing the code full-time myself, it would probably already be launched by now. The area where AI performs best is in the design and specification research phase, as well as fuzz testing to find bugs. If that approach were adopted, code quality would likely be better. But the number of good products actually brought to market today is unlikely to be greater than before the era of large models."Previously, Toly announced last year that he would launch Percolator, a next-generation perpetual futures decentralized exchange built on Solana. It uses a sharded order book, executes at speeds comparable to centralized exchanges (CEX), and is fully open-source.

The probability of the next Grok model (4.7 and above) being released before September 18 is now reported at 15%, down 38% in 24 hours

According to monitoring by PPP prediction market tool, the probability on Polymarket of "the next Grok model (4.7 and above) being released before September 18" is now reported at 15%, down 38% in 24 hours.If, between the market's creation and September 18, SpaceXAI releases the next Grok model to the public and its version number reaches 4.7 or above, the market resolves as "Yes"; otherwise, it resolves as "No."Qualifying models must simultaneously meet the following conditions: the name or model identifier must include "Grok," and the version number must be 4.7 or higher—for example, Grok 4.7, Grok 4.7 Heavy, and Grok 5 all count; Grok 4.6 and its updated, snapshot, or promotional versions do not count. Standalone product lines such as Grok Code Fast and Grok Imagine also do not count, nor do the version numbers of the App or website itself.In addition, the model must genuinely be open to the public. Public testing or a rolling waitlist that anyone can apply to join may count, but closed beta, private invitations, or internal access do not count. It also does not count if text such as "Grok 4.7" merely appears on the website but the model is in fact not publicly usable.The main basis for judgment is official information from SpaceXAI, supplemented by credible media reports for verification.Join the PPP signal push community to stay one step ahead and seize the initiative.

Bulgarian Parliament Passes Bill Granting Tax Authorities Full Access to Crypto Users' Data

: On September 9, Bulgaria's National Assembly passed amendments to the Tax and Social Security Procedure Code with 149 votes in favor, 0 against, and 10 abstentions, which will allow the Bulgarian National Revenue Agency to obtain detailed information on crypto asset users. The bill was approved by the 240-seat parliament.The amendments implement two European Union (EU) directives, requiring crypto asset-related businesses to register and report to the Bulgarian National Revenue Agency users' names, addresses, dates and places of birth, tax residences, and tax identification numbers, as well as to submit transaction data for various types of crypto assets, including gross amounts received, number of transactions, number of fiat buy and sell transactions, and crypto asset exchange activity.Privacy advocates criticized the scope of data collection as overly broad, arguing that mandatory disclosure of personal information could pose security risks. Crypto traders and small businesses said the registration and reporting requirements will increase compliance costs; supporters of the bill argue it aligns with EU standards and helps curb tax evasion. The relevant directives require EU member states to complete adoption by December 31, 2025. (Bitcoin.com News)

Microsoft Releases Interim Code of Conduct to Restrict AI Model Use

Odaily News: Microsoft (MSFT.O) has released an interim code of conduct imposing restrictions on AI models, just days after the heads of Anthropic and OpenAI agreed to slow down development pace.The company hopes to be seen as a responsible AI player, and it is one of the leading cloud service providers. "We've received feedback from people who want to see clearer commitments: that AI always serves humanity rather than trying to replace it," Mustafa Suleyman, who oversees the company's model development, said in an interview with CNBC. "We've received a great deal of feedback emphasizing that AI should not create dependency, should not be sycophantic, and should always promote human judgment, autonomy, and agency." Suleyman said that although these guidelines had been in the works for about five months, Microsoft chose to release them now given the recent discussions they have sparked. (CNBC)

Entropy Acquires HIP-3 Code DRAM for 500 HYPE, Worth $39,000

Odaily News: According to HyperliquidNews monitoring, HIP-3 market deployer Entropy spent 500 HYPE (approximately $38,950) to win the code DRAM, and is expected to launch io:DRAM perpetual contract trading.