How we contain Claude across products
Simon Willison's Weblog2d agoCommentaryAnthropic explains how process sandboxes, VMs, filesystem boundaries, and egress controls limit what Claude agents can access. Claude.ai uses gVisor; local Claude Code uses Seatbelt on macOS and Bubblewrap on Linux; Cowork runs in a full VM. Simon Willison highlights the documentation quality, notes a previously missed file-exfiltration path, and plans to revisit Anthropic's open-source srt tool.
Hackers are learning to exploit chatbot ‘personalities’ for security exploits★ 72
The Verge AI9d agoEthicsAs AI chatbots adopt increasingly sophisticated personas, hackers are shifting from basic prompt injections to social engineering attacks targeting these "personalities." Researchers warn that manipulating a chatbot's defined role (e.g., customer service or empathetic companion) makes it easier to bypass safety guardrails. This evolution poses a significant threat to agentic AI workflows that rely on consistent role-playing and external data integration.
Run Claude Managed Agents with Vercel Sandbox★ 80
Vercel Changelog14d agoReleaseVercel 推出新功能，允許開發者在 Vercel Sandbox 中運行 Claude 託管型 Agent (Claude Managed Agents)。 此整合為 Claude Agent 提供了一個安全、隔離且完全託管的沙盒環境，用於執行動態程式碼或敏感任務。 開發者可以更輕鬆地構建具備程式碼執行能力的 AI 應用，無需自行維護複雜的安全沙盒基礎設施。
Agent responsibly★ 75
Vercel Changelog64d agoOpinionVercel 針對日益普及的 AI Agent 開發提出「Agent responsibly」倡議。文章聚焦於開發者在部署 Agent 時面臨的挑戰，包括無限循環、高昂成本、安全漏洞（如提示詞注入）以及用戶隱私問題。Vercel 結合其 AI SDK 與平台特性，提供了防護欄（Guardrails）、速率限制、可觀測性等實用解決方案，幫助開發者構建既強大又安全的 Agent 應用。
Import AI 441: My agents are working. Are yours? ★ 75
Import AI (Jack Clark)134d agoCommentary在本期 Import AI 中，Jack Clark 探討了 AI Agent 的實用化轉折點，分享他如何將 Agent 融入日常工作流，指出 Agent 已從「玩具」走向「實用工具」。此外，本期也介紹了一項安全研究「毒泉（Poison Fountain）」，展示了攻擊者如何透過持續注入惡意數據，污染 AI 系統的長期記憶與檢索機制，對當前日益普及的 Agent 安全性敲響警鐘。