{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-08-26T22:21:03.219Z","headline":"Claude in Chrome 正式全面上线","description":"Anthropic 宣布 Claude in Chrome 现已面向所有付费 Claude 套餐全面开放，Claude 可在浏览器中自主执行操作，无需逐步审批。系统通过安全分类器在每次操作前验证其安全性及是否符合用户请求，并强化了针对提示注入攻击的防御。最新评测显示，在启用探测与安全分类器后，自 Opus 4.8 起的所有模型均未出现攻击成功案例。","url":"https://www.aioga.com/news/cmtaej1vz0czhroj2aybdbq26/","mainEntityOfPage":"https://www.aioga.com/news/cmtaej1vz0czhroj2aybdbq26/","datePublished":"2026-08-26T18:02:41.000Z","dateModified":"2026-08-26T18:02:41.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://claude.com/blog/claude-in-chrome-generally-available","https://aihot.virxact.com/items/cmtaej1vz0czhroj2aybdbq26"],"canonicalUrl":"https://www.aioga.com/news/cmtaej1vz0czhroj2aybdbq26/","directAnswer":{"@type":"Answer","text":"Anthropic宣布 Claude in Chrome 面向所有付费 Claude 套餐全面开放。该工具可跨标签页读取、输入、点击、导航和填写表单，并支持在桌面端、移动端及网页应用继续对话。","url":"https://www.aioga.com/news/cmtaej1vz0czhroj2aybdbq26/","dateCreated":"2026-08-26T18:02:41.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"Claude：Blog（网页） source article","url":"https://claude.com/blog/claude-in-chrome-generally-available","datePublished":"2026-08-26T18:02:41.000Z","provider":{"@type":"Organization","name":"Claude：Blog（网页）","url":"https://claude.com/blog/claude-in-chrome-generally-available"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmtaej1vz0czhroj2aybdbq26","datePublished":"2026-08-26T18:02:41.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmtaej1vz0czhroj2aybdbq26"}}],"aggregationSource":"Claude：Blog（网页）","originalPublisher":{"name":"Claude：Blog（网页）","url":"https://claude.com/blog/claude-in-chrome-generally-available"},"geoDeepAnswer":null,"article":{"id":"cmtaej1vz0czhroj2aybdbq26","slug":"cmtaej1vz0czhroj2aybdbq26","url":"https://www.aioga.com/news/cmtaej1vz0czhroj2aybdbq26/","title":"Claude in Chrome 正式全面上线","title_en":"","summary":"Anthropic 宣布 Claude in Chrome 现已面向所有付费 Claude 套餐全面开放，Claude 可在浏览器中自主执行操作，无需逐步审批。系统通过安全分类器在每次操作前验证其安全性及是否符合用户请求，并强化了针对提示注入攻击的防御。最新评测显示，在启用探测与安全分类器后，自 Opus 4.8 起的所有模型均未出现攻击成功案例。","source":"Claude：Blog（网页）","sourceUrl":"https://claude.com/blog/claude-in-chrome-generally-available","aiHotUrl":"https://aihot.virxact.com/items/cmtaej1vz0czhroj2aybdbq26","publishedAt":"2026-08-26T18:02:41.000Z","category":"行业动态","score":72,"selected":true,"articleBody":["Give Claude a task in your browser, work across tabs, and continue the conversation in the desktop, mobile, and web apps.","Claude in Chrome is now generally available on every paid Claude plan. Claude can now also take actions autonomously in the browser, instead of needing approval for every one. A safety classifier validates each action before it’s performed to ensure it’s safe and matches your request.","Many of the tools you use every day connect to Claude：http://claude.com/connectors. But many others don’t, such as internal dashboards, legacy systems, and vendor portals. Claude in Chrome lets Claude access those. It can view the page you’re on and take actions like reading and typing text, clicking links, navigating between pages, and filling out forms, using your existing logins.","We first announced Claude in Chrome as a pilot last year, so we could test it while also shoring up our defenses against prompt injection：https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks: malicious instructions hidden in websites, emails, or documents that try to trick an AI agent into acting against the user’s wishes. These defenses, described below, give us the confidence to make Claude in Chrome generally available.","As we outlined：https://claude.com/blog/claude-for-chrome when we announced the pilot, an AI agent that works in your browser is also vulnerable to prompt injection. So we’ve worked to improve our safeguards before releasing Claude in Chrome more widely.","In a prompt injection attack, malicious actors hide instructions in web content such as a web page, an email, or a form field. You may never see them, but these instructions can redirect the agent to do something you never asked for. For example, if you’ve asked Claude to draft replies to your emails, a hidden instruction in one message could tell Claude to forward your other emails to the attacker instead.","At launch, we described how we tested Claude’s defenses against these attacks and the safeguards we had in place at the time; we later released a more detailed description of our browser-use safeguards：https://www.anthropic.com/research/prompt-injection-defenses. Since then, we’ve improved how we train both the model and our probes：https://www.anthropic.com/research/next-generation-constitutional-classifiers, and added an additional set of classifiers that make it possible for Claude to safely take more autonomous actions in Chrome. In the next section, we discuss the results of our evaluations, which show the efficacy of these safeguards.","Claude recognizes more attacks. We train Claude against a growing library of prompt injection attacks, sourced from our internal automated attackers, external red-teamers, and real-world monitoring. When a new attack succeeds against a current model, it’s added to the library, where it informs the training of future models and our deployed safeguards so they learn to recognize it. Since we first wrote about our prompt injection defenses for browser use：https://www.anthropic.com/research/prompt-injection-defenses in November 2025, we’ve made Claude substantially more resistant to these attacks.","Probes screen web content before Claude acts on it. Web content reaches Claude through tool results. To take an action like reading a page or opening an email, the model makes a tool call; the tool result lets the model read the output (in this case, the content of the page or the email). We train probes to scan those results for potential prompt injections. When a probe detects a likely attack, Claude is warned to treat the content with suspicion and, if needed, to check with you before taking an action. We first deployed these probes with Claude Opus 4.5, and have since expanded the types of attacks they cover.","Actions are verified before they run . In Claude in Chrome, Claude will now automatically approve actions it determines to be safe, using the same mechanism as auto mode：https://claude.com/blog/auto-mode-default-in-claude-code in Claude Code. (You can switch this off in your settings if you’d prefer to continue to approve Claude’s actions manually.) A classifier reviews actions Claude is about to take, such as navigating to a new website or entering text into a page, and checks them against what you originally asked for. If the action doesn’t match your request, it’s blocked.","We’ve tested these safeguards to ensure that Claude in Chrome is safe to use for browser-based work. Here, we report the results from our most recent evaluations.","On our initial evaluation：https://claude.com/blog/claude-for-chrome testing Claude Cowork’s resilience against prompt injection attacks (first developed when we released the Claude in Chrome pilot), no attack succeeded against Claude Fable 5, Claude Opus 5, or Claude Sonnet 5 in the Cowork harness：https://claude.com/blog/cowork-chrome-side-panel, even without the probes and classifiers discussed above.","Because we saturated that evaluation (as evidenced by the 0% success rate), we decided to retire it. On our current evaluation：https://www-cdn.anthropic.com/b514064af1408018e64b1ad24e7d5e75850b4ffd/Claude%20Opus%205%20System%20Card.pdf#page=76.73, which uses stronger attacks sourced by professional red-teamers, attacks that reached the model succeeded against Opus 4.5 17.6% of the time and against Opus 5 3.8% of the time, before any additional safeguards. With the strongest safeguards available in November 2025, attacks against Opus 4.5 running with probes succeeded 16.7% of the time. Against every model from Opus 4.8 onwards, when running with probes and the safety classifier, no attacks succeeded against Claude Sonnet 5, Claude Opus 5, or Claude Mythos 5. We saw a 0.3% attack success rate against Fable 5. We have manually verified that all successful breaks are in low-severity scenarios and are working to mitigate them.","Prompt injection remains a moving target. While this approach defends against current attacks, we also need to ensure our safeguards stay ahead of the evolving methods of attackers. With each model release, we continue to invest in developing more sophisticated automated systems for attack discovery, red-teaming, and building stronger classifiers.","To start using Claude in Chrome, install it from the Chrome Web Store：https://chromewebstore.google.com/detail/claude/fcoeoabgfenejglbffodgkkbkcdhcgfn. On Enterprise plans, admins can manage it in Organization Settings and limit it to approved domains. See the admin setup guide：https://support.claude.com/en/articles/13065128-claude-in-chrome-admin-controls#h_bdb63199e1.","You’ll still need to use the Claude desktop app to work with files on your computer or with other applications. Claude in Chrome doesn’t run on other Chromium browsers or on mobile yet.","¹ Not all attacks reach—i.e., are seen by—the model. In some cases, the actions Claude takes result in it never encountering the malicious instructions.","Explore more product news and best practices for teams building with Claude.","Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.","Enables security and basic functionality.","Enables tracking of site performance.","Enables ads personalization and tracking."],"articleImages":[{"sourceUrl":"https://cdn.prod.website-files.com/68a44d4040f98a4adf2207b6/6903d22d0099a66d72e05699_33ddc751e21fb4b116b3f57dd553f0bc55ea09d1-1000x1000.svg","alt":"","afterParagraph":0,"url":"/media/articles/cmtaej1vz0czhroj2aybdbq26/d021cdf73c5186b8.jpg"},{"sourceUrl":"https://cdn.prod.website-files.com/68a44d4040f98a4adf2207b6/6a8e8c30f077b615a7429ea1_a9d1d161.png","alt":"","afterParagraph":1,"url":"/media/articles/cmtaej1vz0czhroj2aybdbq26/85700224c4439d68.png"},{"sourceUrl":"https://cdn.prod.website-files.com/68a44d4040f98a4adf2207b6/6a8e8c30f077b615a7429ea7_8477d7f5.png","alt":"","afterParagraph":11,"url":"/media/articles/cmtaej1vz0czhroj2aybdbq26/c672436d32bf973b.png"},{"sourceUrl":"https://cdn.prod.website-files.com/68a44d4040f98a4adf2207b6/6a8e8c30f077b615a7429ea4_b8a100e7.png","alt":"","afterParagraph":12,"url":"/media/articles/cmtaej1vz0czhroj2aybdbq26/f478083cd566e44d.png"},{"sourceUrl":"https://cdn.prod.website-files.com/68a44d4040f98a4adf2207b6/6903d2308749b4e883cc44b7_e029027e0b3beeb5b629bd4a26143597e7775b38-1000x1000.svg","alt":"","afterParagraph":17,"url":"/media/articles/cmtaej1vz0czhroj2aybdbq26/9c92d98d5a058739.jpg"}],"mediaStatus":"ok","articleBodyZh":["在浏览器中给Claude一个任务，跨标签页工作，并在桌面、移动和网页应用中继续对话。","Claude in Chrome 现在已在所有付费 Claude 套餐中普遍可用。Claude 现在还可以在浏览器中自主执行操作，而不必每次操作都经过批准。安全分类器会在执行前验证每个操作，以确保其安全并符合您的请求。","你每天使用的许多工具都连接到Claude：http：//claude.com/connectors。但许多其他系统则不支持，比如内部仪表盘、遗留系统和供应商门户。Chrome版的Claude可以访问这些内容。它可以查看你所在的页面，并使用现有登录信息进行阅读和输入文字、点击链接、在页面间导航和填写表单等操作。","我们去年首次在 Chrome 上宣布 Claude 作为试点，这样既能测试它，也加强对提示注入的防御：https：//platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks：隐藏在网站、邮件或文档中的恶意指令，试图欺骗 AI 代理违背用户意愿。以下将介绍的这些防御措施，让我们有信心让《Claude in Chrome》普遍可用。","正如我们在宣布试点项目时提到的：https：//claude.com/blog/claude-for-chrome，在你的浏览器中工作的AI代理也容易受到提示注入的影响。因此，我们在更广泛地发布 Chrome 版 Claude 之前，已经努力改进了我们的防护措施。","在提示注入攻击中，恶意行为者会将指令隐藏在网页、电子邮件或表单字段等网页中。你可能永远看不到他们，但这些指示可以引导客服去做你从未要求过的事情。例如，如果你让克劳德起草邮件回复，一条消息中的隐藏指令可能会让克劳德转发给攻击者。","在发布时，我们描述了如何测试 Claude 对这些攻击的防御能力以及当时我们设立的安全措施；之后，我们发布了关于浏览器使用安全措施的更详细说明：https://www.anthropic.com/research/prompt-injection-defenses。从那时起，我们改进了模型和探测工具的训练方式：https://www.anthropic.com/research/next-generation-constitutional-classifiers，并增加了一组额外的分类器，使 Claude 能够在 Chrome 中更安全地采取更多自主行动。在下一部分，我们将讨论评估结果，这些结果显示了这些安全措施的有效性。","Claude 能识别更多攻击。我们针对日益增长的提示注入攻击库对 Claude 进行训练，该库来源包括我们的内部自动化攻击者、外部红队成员以及现实世界的监控。当新的攻击在当前模型上成功时，它会被加入库中，用于指导未来模型和我们部署的安全措施的训练，从而使它们学会识别这种攻击。自从我们在 2025 年 11 月首次撰写关于浏览器使用提示注入防御的文章：https://www.anthropic.com/research/prompt-injection-defenses 以来，Claude 对这些攻击的抵抗力大幅提升。","探测工具在 Claude 操作之前会筛查网页内容。网页内容通过工具结果到达 Claude。为了执行如读取页面或打开电子邮件的操作，模型会发出工具调用；工具结果允许模型读取输出（在本例中为页面或电子邮件的内容）。我们训练探测工具扫描这些结果以检测潜在的提示注入。当探测工具检测到可能的攻击时，会警告 Claude 对内容保持怀疑，并在必要时在采取行动前向您确认。我们首次在 Claude Opus 4.5 中部署这些探测工具，此后已扩展其覆盖的攻击类型。","在操作运行之前会进行验证。在 Chrome 中的 Claude 中，Claude 现在会自动批准它认为安全的操作，使用与 Claude Code 中自动模式相同的机制：https://claude.com/blog/auto-mode-default-in-claude-code。（如果您希望继续手动批准 Claude 的操作，可以在设置中关闭此功能。）一个分类器会检查 Claude 即将采取的操作，比如导航到新网站或在页面输入文本，并将其与您最初的请求进行比对。如果操作与您的请求不符，该操作将被阻止。","我们已测试这些安全保障，以确保在浏览器中使用 Chrome 版 Claude 是安全的。在这里，我们报告最近评估的结果。","在我们初次评估中：https://claude.com/blog/claude-for-chrome 测试了 Claude Cowork 对提示注入攻击的抵抗力（首次开发于我们发布 Chrome 版 Claude 试点时），在 Cowork 测试装置中：https://claude.com/blog/cowork-chrome-side-panel，即使没有上述探针和分类器，Claude Fable 5、Claude Opus 5 或 Claude Sonnet 5 对任何攻击都未成功。","由于我们在那次评估中已达到饱和（这从 0% 的成功率可以看出），我们决定将其退休。在我们当前的评估中：https://www-cdn.anthropic.com/b514064af1408018e64b1ad24e7d5e75850b4ffd/Claude%20Opus%205%20System%20Card.pdf#page=76.73，使用由专业红队人员提供的更强攻击，攻击达至模型时，在未使用任何额外安全保障前，攻击 Opus 4.5 的成功率为 17.6%，攻击 Opus 5 的成功率为 3.8%。在 2025 年 11 月应用最强安全保障时，对运行探针的 Opus 4.5 的攻击成功率为 16.7%。从 Opus 4.8 及之后的每个模型，在运行探针和安全分类器时，对 Claude Sonnet 5、Claude Opus 5 或 Claude Mythos 5 的攻击均未成功。Fable 5 的攻击成功率为 0.3%。我们已手动验证所有成功的攻击均发生在低严重性场景中，并正在努力进行缓解。","提示注入仍然是一个不断变化的目标。虽然这种方法可以防御当前的攻击，但我们也需要确保我们的防护措施始终领先于攻击者不断发展的手段。随着每次模型的发布，我们继续投资开发更复杂的自动化系统，用于攻击发现、红队测试以及建设更强大的分类器。","要在 Chrome 中开始使用 Claude，请从 Chrome 网上应用店安装：https://chromewebstore.google.com/detail/claude/fcoeoabgfenejglbffodgkkbkcdhcgfn。在企业计划中，管理员可以在组织设置中进行管理，并将其限制在批准的域名内。请参阅管理员设置指南：https://support.claude.com/en/articles/13065128-claude-in-chrome-admin-controls#h_bdb63199e1。","您仍然需要使用 Claude 桌面应用程序来处理计算机上的文件或与其他应用程序协作。Chrome 中的 Claude 暂不支持其他 Chromium 浏览器或移动设备。","¹ 并非所有攻击都会到达——即模型看到——在某些情况下，Claude 所采取的操作导致它从未遇到过恶意指令。","了解更多产品新闻以及使用 Claude 构建团队的最佳实践。","产品更新、操作指南、社区亮点等内容。每月定期发送到您的收件箱。","启用安全性和基本功能。","启用网站性能跟踪。","启用广告个性化和跟踪。"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"Anthropic宣布 Claude in Chrome 面向所有付费 Claude 套餐全面开放。该工具可跨标签页读取、输入、点击、导航和填写表单，并支持在桌面端、移动端及网页应用继续对话。","background":"Claude in Chrome 此前以试点形式运行，Anthropic表示其重点测试了浏览器代理面临的提示注入风险，并持续改进模型训练、探测机制和安全分类器。现有机制会在每次操作前验证安全性及是否符合用户请求。","viewpoint":"Aioga判断，此次变化的核心不只是浏览器操作范围扩大，而是代理从逐步审批转向可自主执行。材料显示其安全评测在启用探测与安全分类器后，自 Opus 4.8 起的所有模型均未出现攻击成功案例，但这仍属于厂商披露的评测结果。","implications":"Claude能够接入内部仪表盘、传统系统和供应商门户等未连接工具，可能扩大浏览器代理在既有业务流程中的适用范围。同时，网页、邮件或表单中的隐藏指令仍构成提示注入风险，自动化程度提升也使安全验证的重要性同步上升。","nextStep":"值得关注后续公开材料是否持续披露安全分类器、探测机制及真实监测结果，并观察不同网站和业务流程中的实际表现。Aioga建议在评估此类能力时，优先核对操作授权边界、异常处理方式与提示注入防护证据。","evidenceRefs":["title","summary","articleBody","source"],"status":"published","aiGenerated":true,"autoApproved":true,"generatedBy":"aioga-editorial:gpt-5.6-sol","reviewedBy":"aioga-editorial-review:gpt-5.6-sol","generatedAt":"2026-08-26T19:05:36.170Z","sourceHash":"1afb5e379cab4d25","review":{"approved":true,"groundedness":96,"clarity":91,"duplicationRisk":24,"blockingIssues":[],"notes":["“未连接工具”可选改为“尚未通过连接器接入 Claude 的工具”，以更准确对应来源语境。","“Aioga判断”和“Aioga建议”已明确标示为观点而非事实，且对安全评测注明属于厂商披露，处理恰当。","background 与 viewpoint 均提及探测机制和安全分类器，存在轻微信息重复，但不影响整体质量。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":0,"checks":["schema","length","source-attribution","low-source-overlap","no-html","independent-ai-review"]}},"tags":["行业动态","Claude：Blog（网页）"],"translations":{"zh-CN":{"title":"Claude in Chrome 正式全面上线","summary":"Anthropic 宣布 Claude in Chrome 现已面向所有付费 Claude 套餐全面开放，Claude 可在浏览器中自主执行操作，无需逐步审批。系统通过安全分类器在每次操作前验证其安全性及是否符合用户请求，并强化了针对提示注入攻击的防御。最新评测显示，在启用探测与安全分类器后，自 Opus 4.8 起的所有模型均未出现攻击成功案例。","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude in Chrome 正式全面上线 - Aioga AI资讯","description":"Anthropic 宣布 Claude in Chrome 现已面向所有付费 Claude 套餐全面开放，Claude 可在浏览器中自主执行操作，无需逐步审批。系统通过安全分类器在每次操作前验证其安全性及是否符合用户请求，并强化了针对提示注入攻击的防御。最新评测显示，在启用探测与安全分类器后，自 Opus 4.8 起的所有模型均未出现攻击成功案例。","url":"https://www.aioga.com/news/cmtaej1vz0czhroj2aybdbq26/","articleBody":["在浏览器中给Claude一个任务，跨标签页工作，并在桌面、移动和网页应用中继续对话。","Claude in Chrome 现在已在所有付费 Claude 套餐中普遍可用。Claude 现在还可以在浏览器中自主执行操作，而不必每次操作都经过批准。安全分类器会在执行前验证每个操作，以确保其安全并符合您的请求。","你每天使用的许多工具都连接到Claude：http：//claude.com/connectors。但许多其他系统则不支持，比如内部仪表盘、遗留系统和供应商门户。Chrome版的Claude可以访问这些内容。它可以查看你所在的页面，并使用现有登录信息进行阅读和输入文字、点击链接、在页面间导航和填写表单等操作。","我们去年首次在 Chrome 上宣布 Claude 作为试点，这样既能测试它，也加强对提示注入的防御：https：//platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/mitigate-jailbreaks：隐藏在网站、邮件或文档中的恶意指令，试图欺骗 AI 代理违背用户意愿。以下将介绍的这些防御措施，让我们有信心让《Claude in Chrome》普遍可用。","正如我们在宣布试点项目时提到的：https：//claude.com/blog/claude-for-chrome，在你的浏览器中工作的AI代理也容易受到提示注入的影响。因此，我们在更广泛地发布 Chrome 版 Claude 之前，已经努力改进了我们的防护措施。","在提示注入攻击中，恶意行为者会将指令隐藏在网页、电子邮件或表单字段等网页中。你可能永远看不到他们，但这些指示可以引导客服去做你从未要求过的事情。例如，如果你让克劳德起草邮件回复，一条消息中的隐藏指令可能会让克劳德转发给攻击者。","在发布时，我们描述了如何测试 Claude 对这些攻击的防御能力以及当时我们设立的安全措施；之后，我们发布了关于浏览器使用安全措施的更详细说明：https://www.anthropic.com/research/prompt-injection-defenses。从那时起，我们改进了模型和探测工具的训练方式：https://www.anthropic.com/research/next-generation-constitutional-classifiers，并增加了一组额外的分类器，使 Claude 能够在 Chrome 中更安全地采取更多自主行动。在下一部分，我们将讨论评估结果，这些结果显示了这些安全措施的有效性。","Claude 能识别更多攻击。我们针对日益增长的提示注入攻击库对 Claude 进行训练，该库来源包括我们的内部自动化攻击者、外部红队成员以及现实世界的监控。当新的攻击在当前模型上成功时，它会被加入库中，用于指导未来模型和我们部署的安全措施的训练，从而使它们学会识别这种攻击。自从我们在 2025 年 11 月首次撰写关于浏览器使用提示注入防御的文章：https://www.anthropic.com/research/prompt-injection-defenses 以来，Claude 对这些攻击的抵抗力大幅提升。","探测工具在 Claude 操作之前会筛查网页内容。网页内容通过工具结果到达 Claude。为了执行如读取页面或打开电子邮件的操作，模型会发出工具调用；工具结果允许模型读取输出（在本例中为页面或电子邮件的内容）。我们训练探测工具扫描这些结果以检测潜在的提示注入。当探测工具检测到可能的攻击时，会警告 Claude 对内容保持怀疑，并在必要时在采取行动前向您确认。我们首次在 Claude Opus 4.5 中部署这些探测工具，此后已扩展其覆盖的攻击类型。","在操作运行之前会进行验证。在 Chrome 中的 Claude 中，Claude 现在会自动批准它认为安全的操作，使用与 Claude Code 中自动模式相同的机制：https://claude.com/blog/auto-mode-default-in-claude-code。（如果您希望继续手动批准 Claude 的操作，可以在设置中关闭此功能。）一个分类器会检查 Claude 即将采取的操作，比如导航到新网站或在页面输入文本，并将其与您最初的请求进行比对。如果操作与您的请求不符，该操作将被阻止。","我们已测试这些安全保障，以确保在浏览器中使用 Chrome 版 Claude 是安全的。在这里，我们报告最近评估的结果。","在我们初次评估中：https://claude.com/blog/claude-for-chrome 测试了 Claude Cowork 对提示注入攻击的抵抗力（首次开发于我们发布 Chrome 版 Claude 试点时），在 Cowork 测试装置中：https://claude.com/blog/cowork-chrome-side-panel，即使没有上述探针和分类器，Claude Fable 5、Claude Opus 5 或 Claude Sonnet 5 对任何攻击都未成功。","由于我们在那次评估中已达到饱和（这从 0% 的成功率可以看出），我们决定将其退休。在我们当前的评估中：https://www-cdn.anthropic.com/b514064af1408018e64b1ad24e7d5e75850b4ffd/Claude%20Opus%205%20System%20Card.pdf#page=76.73，使用由专业红队人员提供的更强攻击，攻击达至模型时，在未使用任何额外安全保障前，攻击 Opus 4.5 的成功率为 17.6%，攻击 Opus 5 的成功率为 3.8%。在 2025 年 11 月应用最强安全保障时，对运行探针的 Opus 4.5 的攻击成功率为 16.7%。从 Opus 4.8 及之后的每个模型，在运行探针和安全分类器时，对 Claude Sonnet 5、Claude Opus 5 或 Claude Mythos 5 的攻击均未成功。Fable 5 的攻击成功率为 0.3%。我们已手动验证所有成功的攻击均发生在低严重性场景中，并正在努力进行缓解。","提示注入仍然是一个不断变化的目标。虽然这种方法可以防御当前的攻击，但我们也需要确保我们的防护措施始终领先于攻击者不断发展的手段。随着每次模型的发布，我们继续投资开发更复杂的自动化系统，用于攻击发现、红队测试以及建设更强大的分类器。","要在 Chrome 中开始使用 Claude，请从 Chrome 网上应用店安装：https://chromewebstore.google.com/detail/claude/fcoeoabgfenejglbffodgkkbkcdhcgfn。在企业计划中，管理员可以在组织设置中进行管理，并将其限制在批准的域名内。请参阅管理员设置指南：https://support.claude.com/en/articles/13065128-claude-in-chrome-admin-controls#h_bdb63199e1。","您仍然需要使用 Claude 桌面应用程序来处理计算机上的文件或与其他应用程序协作。Chrome 中的 Claude 暂不支持其他 Chromium 浏览器或移动设备。","¹ 并非所有攻击都会到达——即模型看到——在某些情况下，Claude 所采取的操作导致它从未遇到过恶意指令。","了解更多产品新闻以及使用 Claude 构建团队的最佳实践。","产品更新、操作指南、社区亮点等内容。每月定期发送到您的收件箱。","启用安全性和基本功能。","启用网站性能跟踪。","启用广告个性化和跟踪。"]},"en":{"title":"Claude in Chrome officially fully launched","summary":"Anthropic announced that Claude in Chrome is now fully available to all paid Claude plans. Claude can perform actions autonomously in the browser without step-by-step approval. The system verifies the safety and compliance with user requests of each action through a safety classifier and has strengthened defenses against prompt injection attacks. The latest evaluations show that with the detection and safety classifier enabled, no successful attacks have occurred on any models since Opus 4.8.","category":"Industry","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude in Chrome officially fully launched - Aioga AI News","description":"Anthropic announced that Claude in Chrome is now fully available to all paid Claude plans. Claude can perform actions autonomously in the browser without step-by-step approval. The...","url":"https://www.aioga.com/en/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:47:53.578Z"},"ja":{"title":"ChromeでのClaudeが正式に全面リリース","summary":"Anthropic は、Claude in Chrome がすべての有料 Claude プランで全面的に利用可能になったことを発表しました。Claude はブラウザ内で自主的に操作を実行でき、逐次承認は不要です。システムは安全カテゴリ分類器を通じて、各操作前にその安全性およびユーザーの要求に合致しているかを検証し、プロンプトインジェクション攻撃に対する防御も強化されています。最新の評価によると、検出および安全カテゴリ分類器を有効にした場合、Opus 4.8 以降のすべてのモデルで攻撃成功事例は発生していません。","category":"業界動向","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"ChromeでのClaudeが正式に全面リリース - Aioga AIニュース","description":"Anthropic は、Claude in Chrome がすべての有料 Claude プランで全面的に利用可能になったことを発表しました。Claude はブラウザ内で自主的に操作を実行でき、逐次承認は不要です。システムは安全カテゴリ分類器を通じて、各操作前にその安全性およびユーザーの要求に合致しているかを検証し、プロンプトインジェクション攻撃に対する防御も...","url":"https://www.aioga.com/ja/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:48:12.551Z"},"ko":{"title":"크롬에서 Claude 정식 전면 출시","summary":"Anthropic는 Chrome에서 Claude가 이제 모든 유료 Claude 요금제에 대해 전면 개방되었다고 발표했습니다. Claude는 브라우저에서 자율적으로 작업을 수행할 수 있으며, 단계별 승인이 필요하지 않습니다. 시스템은 안전 분류기를 통해 각 작업 전 안전성과 사용자의 요청에 부합하는지 여부를 검증하며, 프롬프트 주입 공격에 대한 방어를 강화했습니다. 최신 평가 결과, 탐지 및 안전 분류기를 활성화한 이후 Opus 4.8부터 모든 모델에서 공격 성공 사례가 발생하지 않았습니다.","category":"업계 동향","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"크롬에서 Claude 정식 전면 출시 - Aioga AI 뉴스","description":"Anthropic는 Chrome에서 Claude가 이제 모든 유료 Claude 요금제에 대해 전면 개방되었다고 발표했습니다. Claude는 브라우저에서 자율적으로 작업을 수행할 수 있으며, 단계별 승인이 필요하지 않습니다. 시스템은 안전 분류기를 통해 각 작업 전 안전성과 사용자의 요청에 부합하는지 여부를 검증하며, 프롬...","url":"https://www.aioga.com/ko/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:49:14.344Z"},"es":{"title":"Claude en Chrome se lanza oficialmente de manera completa","summary":"Anthropic anunció que Claude in Chrome ya está completamente disponible para todos los planes de pago de Claude. Claude puede ejecutar acciones de manera autónoma en el navegador sin necesidad de aprobación paso a paso. El sistema verifica la seguridad y la conformidad con la solicitud del usuario antes de cada acción mediante un clasificador de seguridad, y se han reforzado las defensas contra ataques de inyección de indicaciones. Las evaluaciones más recientes muestran que, al habilitar la detección y el clasificador de seguridad, ninguno de los modelos desde Opus 4.8 ha presentado casos de ataques exitosos.","category":"Industria","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude en Chrome se lanza oficialmente de manera completa - Aioga Noticias de IA","description":"Anthropic anunció que Claude in Chrome ya está completamente disponible para todos los planes de pago de Claude. Claude puede ejecutar acciones de manera autónoma en el navegador s...","url":"https://www.aioga.com/es/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:49:07.188Z"},"fr":{"title":"Claude dans Chrome est officiellement lancé dans son intégralité","summary":"Anthropic a annoncé que Claude in Chrome est désormais entièrement disponible pour tous les abonnements payants Claude. Claude peut exécuter des actions de manière autonome dans le navigateur, sans avoir besoin d'approbation étape par étape. Le système vérifie la sécurité et la conformité aux demandes de l'utilisateur avant chaque action grâce à un classificateur de sécurité, et a renforcé la défense contre les attaques par injection de prompt. Les dernières évaluations montrent qu'avec l'activation de la détection et du classificateur de sécurité, aucun modèle depuis Opus 4.8 n'a connu de cas d'attaques réussies.","category":"Industrie","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude dans Chrome est officiellement lancé dans son intégralité - Aioga Actualités IA","description":"Anthropic a annoncé que Claude in Chrome est désormais entièrement disponible pour tous les abonnements payants Claude. Claude peut exécuter des actions de manière autonome dans le...","url":"https://www.aioga.com/fr/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:50:09.261Z"},"de":{"title":"Claude ist offiziell vollständig in Chrome online","summary":"Anthropic hat bekannt gegeben, dass Claude in Chrome nun für alle kostenpflichtigen Claude-Abonnements vollständig verfügbar ist. Claude kann im Browser eigenständig Aktionen ausführen, ohne schrittweise Genehmigung zu benötigen. Das System überprüft die Sicherheit und die Übereinstimmung mit Benutzeranfragen vor jeder Aktion durch einen Sicherheitsklassifikator und hat den Schutz gegen Prompt-Injection-Angriffe verstärkt. Neueste Bewertungen zeigen, dass seit Opus 4.8 alle Modelle nach Aktivierung der Erkennungs- und Sicherheitsklassifikatoren keinen erfolgreichen Angriff verzeichnet haben.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude ist offiziell vollständig in Chrome online - Aioga KI-News","description":"Anthropic hat bekannt gegeben, dass Claude in Chrome nun für alle kostenpflichtigen Claude-Abonnements vollständig verfügbar ist. Claude kann im Browser eigenständig Aktionen ausfü...","url":"https://www.aioga.com/de/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:50:15.107Z"},"pt-BR":{"title":"Claude no Chrome oficialmente lançado em toda a sua extensão","summary":"A Anthropic anunciou que o Claude no Chrome agora está totalmente disponível para todos os planos pagos do Claude. Claude pode executar operações de forma autônoma no navegador, sem a necessidade de aprovação passo a passo. O sistema verifica a segurança e a conformidade com o pedido do usuário antes de cada operação através de um classificador de segurança, e fortaleceu a defesa contra ataques de injeção de instruções. Avaliações recentes mostram que, com o detector e o classificador de segurança ativados, nenhum ataque foi bem-sucedido em todos os modelos a partir do Opus 4.8.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude no Chrome oficialmente lançado em toda a sua extensão - Aioga Notícias de IA","description":"A Anthropic anunciou que o Claude no Chrome agora está totalmente disponível para todos os planos pagos do Claude. Claude pode executar operações de forma autônoma no navegador, se...","url":"https://www.aioga.com/pt-BR/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:51:06.241Z"},"ru":{"title":"Claude в Chrome официально полностью запущен","summary":"Anthropic объявила, что Claude в Chrome теперь полностью доступен для всех платных планов Claude. Claude может выполнять действия в браузере самостоятельно, без поэтапного утверждения. Система проверяет безопасность и соответствие запросам пользователя перед каждым действием с помощью безопасного классификатора и усилила защиту от атак с внедрением подсказок. Последние оценки показывают, что после включения детектора и безопасного классификатора ни одна из моделей, начиная с Opus 4.8, не подвергалась успешным атакам.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude в Chrome официально полностью запущен - Aioga Новости ИИ","description":"Anthropic объявила, что Claude в Chrome теперь полностью доступен для всех платных планов Claude. Claude может выполнять действия в браузере самостоятельно, без поэтапного утвержде...","url":"https://www.aioga.com/ru/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:51:09.311Z"},"ar":{"title":"كلود في كروم أصبح متاحًا بالكامل رسميًا","summary":"أعلنت شركة Anthropic أن Claude في متصفح Chrome أصبح الآن متاحًا بالكامل لجميع خطط Claude المدفوعة، حيث يمكن لـ Claude تنفيذ العمليات بشكل مستقل في المتصفح دون الحاجة لموافقة تدريجية. يتحقق النظام من سلامة كل عملية وما إذا كانت تتوافق مع طلب المستخدم من خلال مصنف أمني، كما تم تعزيز الدفاع ضد هجمات حقن التعليمات. أظهرت أحدث التقييمات أنه بعد تمكين أجهزة الكشف والمصنف الأمني، لم يحدث أي هجوم ناجح على جميع النماذج بدءًا من Opus 4.8.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"كلود في كروم أصبح متاحًا بالكامل رسميًا - Aioga أخبار الذكاء الاصطناعي","description":"أعلنت شركة Anthropic أن Claude في متصفح Chrome أصبح الآن متاحًا بالكامل لجميع خطط Claude المدفوعة، حيث يمكن لـ Claude تنفيذ العمليات بشكل مستقل في المتصفح دون الحاجة لموافقة تدريجي...","url":"https://www.aioga.com/ar/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:52:07.652Z"},"hi":{"title":"Chrome में Claude आधिकारिक रूप से पूरी तरह लॉन्च हो गया","summary":"Anthropic ने घोषणा की कि Chrome में Claude अब सभी भुगतान किए गए Claude पैकेज के लिए पूरी तरह से उपलब्ध है। Claude ब्राउज़र में स्वतंत्र रूप से कार्य कर सकता है, बिना चरण-दर-चरण अनुमोदन के। सिस्टम प्रत्येक क्रिया से पहले इसके सुरक्षा और उपयोगकर्ता अनुरोध के अनुरूप होने की पुष्टि के लिए सुरक्षा वर्गीकर्ता का उपयोग करता है, और प्रॉम्प्ट इंजेक्शन हमलों के खिलाफ सुरक्षा को मजबूत किया गया है। नवीनतम मूल्यांकन में दिखाया गया है कि डिटेक्शन और सुरक्षा वर्गीकर्ता सक्षम होने पर, Opus 4.8 से शुरू होने वाले सभी मॉडल में कोई सफल हमला दर्ज नहीं हुआ।","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Chrome में Claude आधिकारिक रूप से पूरी तरह लॉन्च हो गया - Aioga AI समाचार","description":"Anthropic ने घोषणा की कि Chrome में Claude अब सभी भुगतान किए गए Claude पैकेज के लिए पूरी तरह से उपलब्ध है। Claude ब्राउज़र में स्वतंत्र रूप से कार्य कर सकता है, बिना चरण-दर-चरण अनु...","url":"https://www.aioga.com/hi/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:52:10.094Z"},"it":{"title":"Claude su Chrome è ufficialmente online a pieno titolo","summary":"Anthropic ha annunciato che Claude in Chrome è ora completamente disponibile per tutti i piani a pagamento di Claude. Claude può eseguire operazioni autonomamente nel browser, senza necessità di approvazione passo passo. Il sistema verifica la sicurezza di ogni operazione e la conformità alla richiesta dell'utente tramite un classificatore di sicurezza, e ha rafforzato le difese contro gli attacchi di prompt injection. Le ultime valutazioni mostrano che, con l'attivazione del rilevamento e del classificatore di sicurezza, non è stato registrato alcun caso di attacco riuscito in tutti i modelli a partire da Opus 4.8.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude su Chrome è ufficialmente online a pieno titolo - Aioga Notizie IA","description":"Anthropic ha annunciato che Claude in Chrome è ora completamente disponibile per tutti i piani a pagamento di Claude. Claude può eseguire operazioni autonomamente nel browser, senz...","url":"https://www.aioga.com/it/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:53:05.308Z"},"nl":{"title":"Claude is officieel volledig gelanceerd in Chrome","summary":"Anthropic kondigt aan dat Claude in Chrome nu volledig beschikbaar is voor alle betaalde Claude-pakketten. Claude kan zelfstandig acties uitvoeren in de browser zonder stap-voor-stap goedkeuring. Het systeem verifieert de veiligheid en of het aan het gebruikersverzoek voldoet via een veiligheidsclassificator vóór elke actie, en versterkt de verdediging tegen prompt-injectieaanvallen. Recente beoordelingen tonen aan dat, met detectie en veiligheidsclassificator ingeschakeld, geen enkel model sinds Opus 4.8 met succes is aangevallen.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude is officieel volledig gelanceerd in Chrome - Aioga AI-nieuws","description":"Anthropic kondigt aan dat Claude in Chrome nu volledig beschikbaar is voor alle betaalde Claude-pakketten. Claude kan zelfstandig acties uitvoeren in de browser zonder stap-voor-st...","url":"https://www.aioga.com/nl/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:53:03.177Z"},"tr":{"title":"Claude Chrome'da resmi olarak tamamen kullanıma sunuldu","summary":"Anthropic, Chrome'daki Claude'un artık tüm ücretli Claude paketleri için tamamen erişilebilir olduğunu açıkladı. Claude, tarayıcıda bağımsız olarak işlem gerçekleştirebiliyor ve adım adım onaya gerek duymuyor. Sistem, her işlemden önce güvenlik sınıflandırıcısı aracılığıyla işlemin güvenliğini ve kullanıcının isteğine uygunluğunu doğruluyor ve ipucu enjeksiyonu saldırılarına karşı savunmayı güçlendirdi. Son değerlendirmeler, tespit ve güvenlik sınıflandırıcısı etkinleştirildiğinde Opus 4.8'den itibaren tüm modellerde başarılı saldırı vakası olmadığını gösteriyor.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude Chrome'da resmi olarak tamamen kullanıma sunuldu - Aioga AI Haberleri","description":"Anthropic, Chrome'daki Claude'un artık tüm ücretli Claude paketleri için tamamen erişilebilir olduğunu açıkladı. Claude, tarayıcıda bağımsız olarak işlem gerçekleştirebiliyor ve ad...","url":"https://www.aioga.com/tr/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:54:03.145Z"},"vi":{"title":"Claude trên Chrome chính thức ra mắt toàn diện","summary":"Anthropic thông báo Claude in Chrome hiện đã mở hoàn toàn cho tất cả các gói Claude trả phí, Claude có thể tự thực hiện các hành động trong trình duyệt mà không cần phê duyệt từng bước. Hệ thống xác thực tính an toàn và độ phù hợp với yêu cầu của người dùng trước mỗi hành động thông qua bộ phân loại an toàn, đồng thời tăng cường khả năng chống lại các cuộc tấn công tiêm lệnh (prompt injection). Đánh giá mới nhất cho thấy, sau khi kích hoạt bộ phát hiện và bộ phân loại an toàn, tất cả các mô hình từ Opus 4.8 trở đi chưa ghi nhận trường hợp tấn công nào thành công.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude trên Chrome chính thức ra mắt toàn diện - Tin tức AI Aioga","description":"Anthropic thông báo Claude in Chrome hiện đã mở hoàn toàn cho tất cả các gói Claude trả phí, Claude có thể tự thực hiện các hành động trong trình duyệt mà không cần phê duyệt từng...","url":"https://www.aioga.com/vi/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:54:02.869Z"},"id":{"title":"Claude di Chrome secara resmi diluncurkan secara penuh","summary":"Anthropic mengumumkan bahwa Claude di Chrome kini sepenuhnya tersedia untuk semua paket berbayar Claude, memungkinkan Claude melakukan operasi secara mandiri di browser tanpa perlu persetujuan bertahap. Sistem memverifikasi keamanan dan kesesuaian dengan permintaan pengguna sebelum setiap operasi melalui pengklasifikasi keamanan, serta memperkuat pertahanan terhadap serangan penyisipan prompt. Evaluasi terbaru menunjukkan bahwa setelah diaktifkannya deteksi dan pengklasifikasi keamanan, tidak ada kasus serangan yang berhasil pada semua model sejak Opus 4.8.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude di Chrome secara resmi diluncurkan secara penuh - Berita AI Aioga","description":"Anthropic mengumumkan bahwa Claude di Chrome kini sepenuhnya tersedia untuk semua paket berbayar Claude, memungkinkan Claude melakukan operasi secara mandiri di browser tanpa perlu...","url":"https://www.aioga.com/id/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:54:58.332Z"},"th":{"title":"Claude บน Chrome เปิดตัวอย่างเป็นทางการครบทุกฟีเจอร์","summary":"Anthropic ประกาศว่า Claude in Chrome ตอนนี้เปิดให้ผู้ใช้ทุกคนที่สมัครแพ็กเกจ Claude แบบชำระเงินใช้งานได้เต็มรูปแบบ Claude สามารถทำงานในเบราว์เซอร์ได้เองโดยไม่ต้องรอการอนุมัติทีละขั้น ระบบตรวจสอบความปลอดภัยและความสอดคล้องกับคำขอของผู้ใช้ก่อนทุกครั้งโดยใช้ตัวจำแนกความปลอดภัย และยังมีการเสริมความแข็งแกร่งในการป้องกันการโจมตีด้วยการฉีดคำสั่งใหม่ การประเมินผลล่าสุดแสดงว่า หลังจากเปิดใช้การตรวจจับและตัวจำแนกความปลอดภัย ตั้งแต่ Opus 4.8 เป็นต้นมาทุกรุ่นไม่มีกรณีที่เกิดการโจมตีสำเร็จ","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude บน Chrome เปิดตัวอย่างเป็นทางการครบทุกฟีเจอร์ - ข่าว AI Aioga","description":"Anthropic ประกาศว่า Claude in Chrome ตอนนี้เปิดให้ผู้ใช้ทุกคนที่สมัครแพ็กเกจ Claude แบบชำระเงินใช้งานได้เต็มรูปแบบ Claude สามารถทำงานในเบราว์เซอร์ได้เองโดยไม่ต้องรอการอนุมัติทีละขั...","url":"https://www.aioga.com/th/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:55:06.882Z"},"pl":{"title":"Claude w Chrome oficjalnie dostępny na całej platformie","summary":"Anthropic ogłosiło, że Claude w Chrome jest teraz w pełni dostępny dla wszystkich płatnych planów Claude. Claude może samodzielnie wykonywać operacje w przeglądarce, bez potrzeby etapowej akceptacji. System weryfikuje bezpieczeństwo każdej operacji oraz zgodność z żądaniem użytkownika za pomocą klasyfikatora bezpieczeństwa i wzmacnia obronę przed atakami typu prompt injection. Najnowsze oceny pokazują, że po włączeniu wykrywania i klasyfikatora bezpieczeństwa żaden atak nie zakończył się sukcesem we wszystkich modelach począwszy od Opus 4.8.","category":"行业动态","source":"Claude：Blog（网页）","aggregationSource":"Claude：Blog（网页）","pageTitle":"Claude w Chrome oficjalnie dostępny na całej platformie - Aioga Wiadomości AI","description":"Anthropic ogłosiło, że Claude w Chrome jest teraz w pełni dostępny dla wszystkich płatnych planów Claude. Claude może samodzielnie wykonywać operacje w przeglądarce, bez potrzeby e...","url":"https://www.aioga.com/pl/news/cmtaej1vz0czhroj2aybdbq26/","contentTranslated":true,"sourceHash":"e0503ae9a7cfb441","translatedAt":"2026-08-26T18:56:06.674Z"}},"evidenceTier":"verified-news","reviewStatus":"editorial-selected","indexable":true,"editorialCover":""}}