Mozilla 週二宣布，Anthropic 的 Claude Mythos AI 模型早期版本在內部測試中識別出 Firefox 瀏覽器中的 271 個安全漏洞，所有漏洞已於本週修復。Mozilla 震驚之際也稱，結果顯示網路安全領域可能正在出現根本性轉變，防禦者或許即將縮小攻擊者長期以來的優勢。
從 22 個到 271 個：Claude Mythos 的安全能力飛躍
Mozilla 此前曾測試另一個 Anthropic 模型，在更早的 Firefox 版本中識別出了 22 個安全敏感漏洞。此次 271 個漏洞的發現，代表了量級上的大幅躍升。
Mozilla 強調，系統發現的所有漏洞「即使是頂尖的人類研究人員」也能發現，且 AI 工具目前尚未揭示人類無法理解的全新漏洞類型。其核心優勢在於大幅加速了這一過程，讓開發者得以在攻擊者利用漏洞之前快速識別。
Claude Mythos 於 2026 年 3 月發布，是 Anthropic 目前最先進的模型，被公司內部資料描述為超越此前 Opus 系列的新型模型，在發布前的測試中，曾在主要作業系統和網路瀏覽器中發現數千個此前未知的漏洞。
Glasswing 計劃：嚴格管控存取的原因
Anthropic 透過「Glasswing 計劃」（Project Glasswing）提供對 Claude Mythos 的有限存取，目前獲准使用的機構僅限於亞馬遜、蘋果、微軟等通過審核的特定科技公司，用途限於軟體漏洞掃描。
這一嚴格管控的背景是：英國 AI 安全研究所測試發現，Claude Mythos 能夠自主執行複雜的網路操作，包括無需人工干預即可完成多階段企業網路攻擊模擬。據知情人士透露，儘管川普政府曾呼籲停止使用 Anthropic 技術，美國國家安全局（NSA）已在機密網路上部署運行 Claude Mythos 預覽版。
雙刃劍：相同能力也可加速網路攻擊
Mozilla 發現的成果具有正反兩面的深遠意義。安全研究人員警告，能夠大規模分析程式碼的 AI 系統可以自動識別廣泛使用的軟體中的可利用漏洞，若落入不法分子手中，將給軟體公司和用戶帶來空前的網路安全威脅，可能催生自動化網路攻擊的新一代形態。
常見問題
Claude Mythos 在 Firefox 找到的 271 個漏洞是什麼類型的問題？
根據 Mozilla 的說明，這些是真實的安全敏感漏洞，「即使是頂尖的人類研究人員」也能發現。Mozilla 表示，AI 工具目前尚未揭示人類無法理解的全新漏洞類型，但其能力在於大規模系統性掃描的速度遠超人工審查，所有問題已於本週完成修復。
Glasswing 計劃的目的是什麼，哪些機構可以使用 Claude Mythos？
Glasswing 計劃是 Anthropic 設立的受控存取計劃，目前僅允許亞馬遜、蘋果、微軟等通過審核的特定科技公司有限使用 Claude Mythos，用途限於軟體安全漏洞掃描。這一限制反映了 Anthropic 對該模型雙重用途風險的高度謹慎。
此次發現對整個網路安全行業有何深遠意義？
Mozilla 表示，AI 工具的出現可能讓防禦者首次有機會縮小攻擊者長期以來的優勢，實現「決定性的勝利」。然而，研究人員同時警告，相同的能力也可被攻擊者利用，加速自動化網路攻擊的規模和效率，因此 AI 安全工具的存取管控至關重要。
Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to
Disclaimer.
Articoli correlati
AI16Z, ELIZAOS Creators Sued Over $2.6B Fraud Allegations; Token Crashes 99.9% From Peak
Federal class action accuses AI16Z/ELIZAOS of a $2.6B crypto fraud via fake AI claims and deceptive marketing, alleging insider favoritism and a staged autonomous system; seeks damages under consumer protection laws.
Abstract: This report covers a SDNY federal class-action filed April 21 accusing AI16Z and its rebrand ELIZAOS of a $2.6 billion crypto fraud involving fake AI claims and deceptive marketing. The suit alleges a manufactured link with Andreessen Horowitz and a non-autonomous system. It details a peak valuation in early 2025, a 99.9% crash, and about 4,000 losing wallets, with insiders receiving ~40% of new tokens. Plaintiffs seek damages and equitable relief under New York and California consumer-protection laws. Regulators in Korea and major exchanges have warned or suspended related trading.
GateNews8m fa
Tencent and Alibaba Negotiate Investment in DeepSeek, Valuation Exceeds $200 Billion
Gate News message, April 22 — Tencent and Alibaba are in negotiations to invest in DeepSeek, according to The Information. DeepSeek's valuation exceeds $200 billion.
GateNews27m fa
Cambricon Expresses Optimism on 10 Billion Yuan Revenue Target Amid AI Chip Demand Surge
Cambricon's 2025 briefing signals growth potential amid rising AI demand; CEO Chen Tianshi stresses independent IP, industry backing, and long-term partnerships to manage competition, inventory, and pending litigation.
GateNews29m fa
Microsoft Unveils AI Agent Commerce Infrastructure: Publisher Marketplace, Merchant Protocols, and Ad Tools
Gate News message, April 22 — Microsoft's AI monetization vice president Tim Frank announced a suite of commercial infrastructure updates designed for the "agentic web" era, enabling publishers, merchants, and advertisers to remain discoverable and tradable as AI agents make purchasing decisions on
GateNews1h fa
NeoCognition Raises $40M in Seed Funding for On-the-Job Learning AI Agents
Gate News message, April 22 — AI research lab NeoCognition announced the completion of a $40 million seed round, emerging from stealth mode. Founded by Ohio State University Associate Professor Yu Su, along with Xiang Deng and Yu Gu, the company is headquartered in Palo Alto, California. The round w
GateNews1h fa
