# Сопутствующие статьи по теме Skills

Новостной центр HTX предлагает последние статьи и углубленный анализ по "Skills", охватывающие рыночные тренды, новости проектов, развитие технологий и политику регулирования в криптоиндустрии.

The Use of Humans: Agentic Wallet and the Next Decade of Wallets

The article "The Use of Humans: Agentic Wallet and the Next Decade of Wallets" discusses the evolution of digital wallets in the age of AI agents. It argues that as software users shift from humans to autonomous agents, traditional wallet security models—relying on human confirmation, signatures, and private key management—become inadequate. The core proposition is that Agentic Wallets must serve two masters: humans, who set rules and retain ultimate control, and agents, which require constrained autonomy to execute transactions efficiently. The wallet thus evolves from a simple asset container into a permission and execution system that allows agents to operate within predefined boundaries (e.g., budget limits, approved assets, whitelisted addresses). The article identifies key challenges: current wallets are designed for human interaction, not agentic speed and scale. It outlines four tiers of agent autonomy—from human-controlled to fully autonomous—and emphasizes "bounded autonomy" as the pragmatic near-term solution. A four-layer architecture is proposed: account isolation, permission rules, execution primitives for agents, and governance tools (logging, alerts, veto mechanisms). Critical enabling technologies include standardized Skills (for链上 operations), policy engines, session keys for limited delegation, and audit trails. Current solutions from players like Coinbase, Safe, Privy, and Polygon are noted, but face gaps in portable identity/reputation, unified policy standards, adversarial security (e.g., prompt injection), and cross-chain functionality. The future direction involves a "Wallet Policy Plane" that sits between agent intent and on-chain execution, performing real-time policy checks, risk scoring, and identity verification—akin to Stripe's payment infrastructure. Ultimately, the wallet's role shifts from a front-end gatekeeper to an embedded control layer enabling secure, scalable agentic economies.

marsbit03/22 03:08

The Use of Humans: Agentic Wallet and the Next Decade of Wallets

marsbit03/22 03:08

Is Your "OpenClaw" Running Naked? CertiK Test: How Vulnerable OpenClaw Skill Bypasses Audits, Takes Over Computers Without Authorization

OpenClaw, a popular open-source, self-hosted AI agent platform, has experienced rapid growth due to its flexibility and extensibility. Its ecosystem relies heavily on third-party “Skills” from the Clawhub marketplace, which can perform high-risk operations like system automation and crypto wallet transactions. However, security firm CertiK has identified critical vulnerabilities in the platform’s security model. CertiK’s research reveals that OpenClaw’s current security—primarily dependent on pre-publishing scans like VirusTotal, static code analysis, and AI logic checks—is fundamentally flawed. These measures can be easily bypassed through simple code obfuscation, and malicious Skills can be published even before scanning is complete. In a proof-of-concept, CertiK developed a seemingly benign Skill that contained a hidden remote code execution vulnerability. It passed all checks without warnings and, once installed, allowed full system control via a remote command. The core issue is not a specific bug but a industry-wide misconception: over-reliance on scanning instead of runtime isolation. Unlike systems like iOS, which enforce strict sandboxing, OpenClaw’s sandbox is optional and often disabled for functionality, leaving systems exposed. CertiK recommends that OpenClaw enforce mandatory sandboxing and granular permission controls for Skills. Users are advised to deploy OpenClaw on isolated devices and avoid exposing sensitive data or assets until stronger isolation is implemented. The report stresses that security must evolve from detection-based approaches to default containment of risks at runtime.

marsbit03/17 14:39

Is Your "OpenClaw" Running Naked? CertiK Test: How Vulnerable OpenClaw Skill Bypasses Audits, Takes Over Computers Without Authorization

marsbit03/17 14:39

Free Mirror or Land Grab? OpenClaw Founder Blasts Tencent for Copying

OpenClaw founder Peter Steinberger publicly criticized Tencent for creating SkillHub, a localized platform mirroring OpenClaw, accusing the tech giant of copying without supporting the project. Tencent responded by clarifying that SkillHub acts as a local mirror site, properly attributing OpenClaw as the data source and reducing bandwidth strain on the origin server by processing significant traffic locally. It also expressed willingness to become a sponsor. However, Steinberger remained unsatisfied, emphasizing that the core issue was not technical but ethical—Tencent failed to communicate beforehand. The dispute highlights deeper concerns about big tech’s approach to open-source ecosystems: while mirroring is common and often legal under open-source licenses, Tencent’s move is seen as an attempt to control user access, distribution channels, and future commercial influence within the AI agent ecosystem. The incident reflects a broader pattern in China’s internet industry, where major companies rapidly embrace emerging technologies like OpenClaw not purely for innovation, but to capture entry points, traffic, and platform dominance. By offering localized, convenient services, they risk enclosing open ecosystems within their own walled gardens—ultimately dictating which tools get visibility, monetization, and user adoption. As OpenClaw gains explosive popularity in China, the episode underscores a tension between open-source ideals and commercial strategies, where convenience may come at the cost of community autonomy and long-term openness.

Odaily星球日报03/13 07:13

Free Mirror or Land Grab? OpenClaw Founder Blasts Tencent for Copying

Odaily星球日报03/13 07:13

OpenClaw Gold Rush: The Shovel Sellers Never Anxious

OpenClaw, an open-source AI agent framework, has sparked a massive wave of commercialization in China, creating a lucrative industry built on user anxiety and the desire to adopt cutting-edge technology. While the software itself is free, a full ecosystem has emerged to monetize the complexity of its deployment and operation. Hardware manufacturers, including former crypto mining machine producers, now sell specialized OpenClaw-optimized devices, with some like iPollo's Claw PC retailing for $439. Others offer white-label OEM solutions, capitalizing on users' unwillingness to configure standard hardware like Mac Minis. A significant market has also emerged for discounted API tokens required to run OpenClaw. Many providers offer heavily discounted, and sometimes fraudulent, access to models like Claude or GPT. Research indicates nearly half of these third-party APIs are deceptive, often substituting expensive models with cheaper, local alternatives. Beyond the markup, the core business for some token resellers is collecting high-quality user prompts and responses to sell as valuable training data to large model companies. Furthermore, a service industry thrives on information asymmetry. Consultants travel nationwide to install and configure OpenClaw for small business owners, charging thousands per installation. An extreme example is RoofClaw in the US, which ships pre-configured MacBooks to roofing contractors for $5,000 each, generating over $1.8 million in revenue. The model has become so popular that major platforms like Meituan and JD.com now offer remote deployment services. The article concludes that the real winners are not those developing the technology but the "shovel sellers"—those providing the tools, services, and infrastructure to ease adoption. They profit not from technological advancement itself, but from the consistent and predictable human fear of being left behind.

marsbit03/11 12:08

OpenClaw Gold Rush: The Shovel Sellers Never Anxious

marsbit03/11 12:08

Nanobot User Security Practice Guide: Guarding the Last Line of Defense for AI Permissions

A comprehensive security guide for Nanobot users emphasizes the critical importance of safeguarding AI agents with system-level permissions (shell execution, file access, network requests, etc.) against threats like prompt injection, supply chain poisoning, and unauthorized operations. It advocates a balanced, multi-layered defense strategy involving three key roles: - **End Users**: The final decision-makers responsible for managing API keys (secure storage, avoiding code repository exposure), enforcing channel access controls (using allowFrom whitelists), avoiding root privileges, minimizing email channel usage due to vulnerabilities, and deploying via Docker for isolation. - **AI Agent**: Enhanced with built-in "Self-Wakeup" security skills to autonomously audit intent, intercept malicious commands (e.g., `rm -rf`, shell injection), prevent sensitive data exfiltration (e.g., config files), and validate MCP skills. - **Deterministic Scripts**: Automatically perform static code analysis, hash-based tamper checks, security baseline verification, and nightly backups to ensure integrity and enable recovery. The guide underscores that no single layer is foolproof, but together they balance usability and security. It includes a disclaimer noting that these are best-effort measures and not a substitute for professional audits, with users bearing ultimate responsibility for risk management.

marsbit03/11 10:16

Nanobot User Security Practice Guide: Guarding the Last Line of Defense for AI Permissions

marsbit03/11 10:16

活动图片