Grok Bot 可免确认直接刷卡的安全质疑
作者引用 xAI 文档指出 Grok Bot 的 Vault 保存卡片后结账可自动扣款、已完成的购买无法撤销，并对比 Claude Code 技能漏洞数据。
THERE'S NO WAY THIS IS SAFE. Grok Bot can pay with your card without asking you first. Straight from xAI's own docs: → save your card in the Bot's Vault → "a card saved in Vault fills in at checkout without a separate approval each time" → hit "Always allow" once and that kind of purchase stops asking → an approval can't undo a purchase that already went through Elon, October 15: "If Grok Bot makes a mistake with a purchase or leaks your card, SpaceX will make you whole." xAI's terms, updated 3 days earlier: you are responsible for "Agentic Actions." Same problem on the coding side. Claude Code skills run with your permissions. NVIDIA scanned 31,132 agent skills: 26.1% had vulnerabilities, 5.2% looked malicious. Agents act as you now. The article below starts with the scan, then the 10 skills I'd actually let run on my machine. Would you give an agent your card?
这条是引用着另一条写的：x.com/kyzorox/status/2111897393041932316