cicy-desktop 2.1.65 → 2.1.67

package/workers/render/src/termsText.js

@@ -0,0 +1,9 @@
+// AUTO-GENERATED from w-10065 terms docs (草案 v0.2). Full terms body for the
+// first-run gate's '查看完整条款' / 'View full terms'. Both locales now have
+// the complete counsel-reviewed text.
+export const TERMS_VERSION = "0.2";
+
+export const TERMS_FULL = {
+  "zh-CN": "# CiCy Desktop 用户协议与授权说明\n\n> 首次启动展示 · 草案 v0.2 · 2026-06-07\n> **本协议为使用本软件的前提:你必须阅读并点击\"同意并继续\"后,方可使用 CiCy Desktop。**\n> （内部草案,非最终法律文本;正式发布前须经中、欧两地持牌律师审定。本文同时作为应用内首启条款的内容源。）\n\n---\n\n## 一眼看懂(你点\"同意\"即代表接受以下要点)\n\n1. **CiCy Desktop 是在你自己电脑上运行的 AI Agent 编排工具**,帮你调度第三方 AI 助手(Claude Code、Codex、Gemini 等)完成工作。\n2. **第三方 AI 与账号是你和对应厂商之间的事**:用哪个 AI、用谁的账号、产生的用量与费用,由你决定并承担,与 CiCy 无关。\n3. **为完成工作,本软件需要一系列对你电脑的授权**(运行程序、执行脚本、控制浏览器、访问文件等)。下面第 3 条逐项讲清,**你对经本软件发起的操作负全部责任**。\n4. **你的数据默认留在本机**,CiCy 不以运营为目的上传你的工作内容。\n5. **两项高风险能力需要你单独、额外授权**,默认关闭:**(a) HTTPS 流量本地审计(需装证书、需管理员)**、**(b) 本机命令/脚本执行**。整体同意本协议**不等于**开启它们。\n6. **严禁**用本软件控制未授权设备、违法多账号/刷量、绕过平台风控、违规抓取等。\n\n> 不接受以上任一条,请点击\"不同意并退出\"。**未同意则无法使用本软件。**\n\n---\n\n## 1. 关于 CiCy Desktop\n\n1.1 CiCy Desktop(下称\"本软件\")是一款在你本地设备运行的 AI Agent 编排与自动化工具,用于组织、调度第三方 AI 命令行/客户端(如 Claude Code、OpenAI Codex、Google Gemini CLI 等,下称\"第三方 Agent\")协同完成任务。\n\n1.2 本软件本身**不提供大模型推理服务**,也不内置任何用于规避地域限制的跨境网络通道。模型能力由你选择的后端提供。\n\n## 2. 第三方 Agent 与账号(免责)\n\n2.1 你可在本软件中接入第三方 Agent。**你与第三方 Agent 及其模型/服务提供商之间的关系(账号注册、登录、授权、用量、计费、合规)由你与该提供商直接建立并承担,与 CiCy 无关。**\n\n2.2 **官方登录责任自负**:若你以第三方提供商的官方账号登录其 Agent,该行为及其后果(包括是否符合该提供商所在地与你所在地的可用性政策)由你自行负责。CiCy 不引导、不代理、不内置任何专用于规避地域限制的网络通道。\n\n2.3 你须遵守你所接入的每一个第三方 Agent 与模型服务的用户协议与使用政策。\n\n---\n\n## 3. 授权说明(本软件需要你授予的权限)\n\n> 为让 AI Agent 替你完成工作,本软件需要下列对你设备的授权。我们按风险高低分为三类,并说明每项的**用途、作用范围、风险、能否撤销**。**透明披露是本协议的核心。**\n\n### 3A. 基础授权(同意本协议即授予,为软件运行所必需)\n\n| 授权 | 用途 | 范围与限制 | 可否撤销 |\n|---|---|---|---|\n| **本地文件读写** | 读写工作区、配置、任务历史 | 限于本软件工作目录及你在任务中明确指向的文件 | 卸载软件即终止 |\n| **运行第三方 Agent** | 在你电脑上启动 claude/codex/gemini 等子进程 | 仅运行你已接入的 Agent | 移除对应 Agent |\n| **本地数据存储** | 保存审计/历史/配置于本机 | 数据留本机,不上传 | 可在设置清除 |\n| **读取系统与窗口信息** | 调度、展示运行状态 | OS/硬件概要、窗口列表;不含文件内容 | 卸载即终止 |\n| **本机网络访问** | 代表你向你选择的模型后端/服务发起请求 | 按你的配置出网;不内置规避地域限制的通道 | 可在设置关闭联网功能 |\n\n### 3B. 高风险授权(默认关闭,需你**单独、额外**同意后才启用)\n\n> 整体同意本协议**不会**自动开启以下任何一项。每项都需要你在应用内单独操作开启,并可随时关闭。\n\n**① 本机命令与脚本执行**\n- **用途**:让 AI Agent 能在你电脑上执行 shell / Python / Node 脚本,以完成自动化任务。\n- **风险(请认真阅读)**:这等同于授予程序在你电脑上**运行代码的能力**。请仅在你信任的任务场景启用。\n- **门控与保护**:仅作用于**本机**;全部执行写入本地审计留痕;对高危命令可设二次确认。\n- **撤销**:可在设置中关闭执行能力。\n\n**② 远程编排配对**\n- **用途**:允许你**显式配对**的编排端(可能运行在云端或另一台设备)通过安全信道驱动本机 Agent,实现远程协作调度。\n- **风险**:配对后,远程编排端可触发上述本机能力。\n- **门控与保护**:**仅连接你显式配对、授权令牌可随时吊销**的编排端;绝不静默安装、绝不自我传播;你可解除配对立即断开。\n- **撤销**:解除配对 / 吊销令牌。\n\n**③ 系统根证书信任库写入(需管理员授权)**\n- **用途**:为下方第 4 条的 HTTPS 流量本地审计所必需——将一张**仅属于本机**的审计证书装入系统信任库。\n- **风险**:写入系统信任根是高敏感操作,故需管理员授权,且本软件**绝不**预置或分发统一证书(详见第 4 条)。\n- **撤销**:`uninstall-ca` 一键卸载并清除同意标记。\n\n**④ HTTPS 流量本地审计** —— 见第 4 条专章。\n\n### 3C. 任务执行中按需使用的授权(可在设置中逐项关闭)\n\n| 授权 | 用途 | 范围与限制 |\n|---|---|---|\n| **浏览器控制** | 打开/导航网页、在你授权的浏览器会话中读取页面或注入脚本、读写 Cookie | 仅作用于本软件管理或你授权的浏览器会话 |\n| **剪贴板读写** | 在自动化流程中传递文本 | 按任务需要;可关闭 |\n| **屏幕/窗口截图** | 让 Agent\"看见\"界面以完成操作 | 截图留本机,不上传 |\n| **多账号隔离 + 每账号代理** | 为不同身份隔离 Cookie/存储/出口 IP(如多店运营、多客户隔离交付) | **仅供合法的多账号管理;不得用于违法养号、刷量或对抗平台风控**(见第 5 条) |\n\n---\n\n## 4. HTTPS 流量本地审计(可选 · 默认关闭 · 需显式同意 · 需管理员)\n\n4.1 本软件提供可选的\"HTTPS 流量本地审计\"功能,用于对**你本机**发往 AI 厂商(如 Claude / OpenAI / DeepSeek / Gemini)的加密请求进行本地安全审计、敏感信息脱敏与策略治理。\n\n4.2 **该功能默认关闭。** 仅当你在应用内主动点击启用、并完成管理员授权后才生效。启用时:\n- 本软件在你本机生成一张**仅属于本机、独一无二**的审计证书(绝不使用预置或跨设备共用的证书),并请求装入本机系统信任库;\n- 安装后,你本机到上述 **AI 厂商域名**的 HTTPS 流量将被本地解密以供审计;**解密范围严格限于这些 AI 厂商域名;其余一切流量(网银、社交、其他网站等)不被解密、不被读取**;\n- **解密后的数据仅保存在你本机**,不上传至任何服务器。\n\n4.3 **可随时撤销**:你可在设置中关闭该功能并卸载证书(`uninstall-ca`),撤销后不再解密任何流量,同意标记一并清除。\n\n4.4 **多用户/团队场景的告知义务**:若你将本软件部署于供多人使用的设备或环境,并对他人流量启用审计,则你(部署方)须自行、事先向相关人员明确告知并取得同意,且仅限经营管理之必要范围。该告知与同意义务由你承担。\n\n## 5. 自动化与设备控制的合法使用(用户责任)\n\n5.1 本软件第 3 条所述能力均为通用自动化/RPA 工具能力。\n\n5.2 **授权边界**:这些能力仅可作用于**你本人拥有或已获明确授权**的设备与账号。**严禁**访问或控制你无权限的计算机系统。\n\n5.3 **禁止用途**,包括但不限于:\n- (a) 入侵、非法控制、干扰他人计算机信息系统;\n- (b) 违反平台规则的批量注册、刷量、刷单、薅羊毛及其他违法多账号行为;\n- (c) 绕过、对抗目标平台的风控或技术保护措施;\n- (d) 违反目标网站 robots 协议、用户条款或法律的数据抓取;\n- (e) 无合法依据抓取或处理他人个人信息;\n- (f) 任何违反你所在地或目标对象所在地法律法规的行为。\n\n5.4 **你对经由本软件发起的全部操作及其后果承担全部责任。** CiCy 仅提供通用工具能力,不对你的具体使用方式负责;若你的使用导致第三方索赔或行政/刑事追究,由你承担并使 CiCy 免责。\n\n## 6. 个人信息与隐私\n\n6.1 本软件遵循本地优先、最小必要原则处理数据,不以运营为目的收集上传你的工作内容。\n\n6.2 当你在涉及个人信息的场景使用本软件(对含个人信息的流量审计、抓取或处理),在适用《个人信息保护法》(中国)、GDPR(欧盟)等法律时,**你或你所属组织可能构成个人信息处理者 / 数据控制者**,须自行履行相应义务(告知同意、处理记录、数据主体权利响应、必要时的数据保护影响评估与泄露通报等)。\n\n6.3 涉及欧盟个人数据出境的,须具备 GDPR 第五章要求的合法传输机制(如标准合同条款 SCC)与你的明确同意。\n\n6.4 **按地区选择模型后端**:为符合不同司法辖区法律,本软件按你所在地区选择默认模型后端(例如中国大陆用户默认使用已在中国境内备案的模型服务);跨境传输个人数据须经你单独明确同意。\n\n## 7. AI 生成内容\n\n7.1 你使用本软件及第三方 Agent 生成的内容,须遵守你所在地关于 AI 生成合成内容标识的法律义务(如中国《人工智能生成合成内容标识办法》、欧盟 AI Act 的透明度要求)。CiCy 在适用范围内提供标识/披露能力,但合规使用责任在你。\n\n## 8. 合规与适用法域\n\n8.1 你须遵守你所在地、以及你的操作所指向对象所在地的全部适用法律、法规与平台条款,**尤其包括中国与欧盟的相关规定**。\n\n8.2 本软件按\"法域感知\"原则对不同地区适用不同默认与限制,但这不免除你自行确保合规的责任。\n\n## 9. 免责声明与责任限制\n\n9.1 本软件按\"现状\"提供,在适用法律允许的最大范围内,不作适销性、特定用途适用性的默示担保。\n\n9.2 在适用法律允许的最大范围内,对于你违反本条款,或因你使用第三方 Agent / 自动化能力 / 审计功能产生的任何直接或间接损失、第三方索赔、行政或刑事责任,CiCy 不承担责任,且你应使 CiCy 免受相关追偿。\n\n9.3 本条款不排除或限制依强制性法律不可排除的责任(如人身伤害或欺诈责任)。\n\n## 10. 同意与生效\n\n10.1 **本协议是使用本软件的前提。** 点击\"同意并继续\",即表示你已阅读、理解并接受本协议全部条款;点击\"不同意并退出\"或不作选择的,**无法使用本软件**。\n\n10.2 第 3B 条高风险授权(本机命令/脚本执行、远程编排配对、根证书写入)与第 4 条 HTTPS 审计,**需你在应用内单独操作开启**;你对本协议的整体同意**不构成**对这些功能的启用。\n\n10.3 你可随时在设置中查看、关闭或撤销第 3 条所列各项授权。\n\n---\n\n> **实施备注(非条款,给产品/法务):**\n> - 首启 = 阻断式条款:展示\"一眼看懂\"6 条 + \"同意并继续 / 不同意并退出\";未同意不进主界面。完整条款可展开/外链。\n> - **第二道独立同意**:第 4 条审计 = w-10026 的同意卡片三态(未授权/已授权/处理中);第 3B① 执行能力、3B② 远程配对建议也各设独立开关与首用确认。\n> - 后端门控:第 4 条对应 w-10084 的 `~/cicy-ai/mitm/ca-trust-consent` flag + `exec cicy-code mitm install-ca/uninstall-ca`。\n> - 容器/无头部署:`CICY_CA_TRUST_CONSENT=1` 逃生门 = \"设置该环境变量即代表部署方代表其环境作出第 4.4 条告知同意\",责任落部署方;须写入部署文档。\n> - 正式版:中文 + 英文(欧盟)双语;按法域分版本可能调整第 3C/6/8 条措辞;高风险授权建议保留**用户可撤销的审计日志**作为同意证据。\n",
+  en: "# CiCy Desktop — Terms of Use & Authorization Notice\n\n> Shown on first launch · Draft v0.2 · 2026-06-07\n> **Acceptance of these Terms is a precondition to using the software: you must read and click \"Agree and continue\" before you can use CiCy Desktop.**\n> (Internal draft, not final legal text; must be reviewed by licensed counsel in both China and the EU before public release. This document also serves as the content source for the in-app first-launch terms.)\n\n---\n\n## At a glance (clicking \"Agree\" means you accept the following)\n\n1. **CiCy Desktop is an AI agent orchestration tool that runs on your own computer**, helping you coordinate third-party AI assistants (Claude Code, Codex, Gemini, etc.) to get work done.\n2. **Third-party AI and accounts are between you and the respective provider**: which AI you use, whose account you log in with, and the resulting usage and fees are your decision and your responsibility — not CiCy's.\n3. **To do the work, the software needs a range of authorizations on your computer** (running programs, executing scripts, controlling the browser, accessing files, etc.). These are explained item by item in Section 3. **You are fully responsible for the operations initiated through the software.**\n4. **Your data stays on your machine by default.** CiCy does not upload your work content for operational purposes.\n5. **Two high-risk capabilities require your separate, additional authorization and are off by default**: **(a) local HTTPS traffic auditing (installs a certificate, requires administrator)** and **(b) local command/script execution**. Accepting these Terms as a whole does **not** turn them on.\n6. **You must not** use the software to control unauthorized devices, conduct unlawful multi-accounting / traffic faking, circumvent platform risk controls, or scrape in violation of terms.\n\n> If you do not accept any of the above, please click \"Decline and exit.\" **Without acceptance, you cannot use the software.**\n\n---\n\n## 1. About CiCy Desktop\n\n1.1 CiCy Desktop (the \"Software\") is an AI agent orchestration and automation tool that runs on your local device, used to organize and coordinate third-party AI command-line tools / clients (such as Claude Code, OpenAI Codex, Google Gemini CLI, etc., the \"Third-Party Agents\") to complete tasks collaboratively.\n\n1.2 The Software itself **does not provide large-model inference services**, nor does it embed any cross-border network channel designed to circumvent regional restrictions. Model capabilities are provided by the backend you choose.\n\n## 2. Third-Party Agents and Accounts (Disclaimer)\n\n2.1 You may connect Third-Party Agents within the Software. **The relationship between you and a Third-Party Agent and its model/service provider (registration, login, authorization, usage, billing, compliance) is established directly between you and that provider and borne by you — it is unrelated to CiCy.**\n\n2.2 **You are solely responsible for official logins.** If you log in to a Third-Party Agent using that provider's official account, that act and its consequences (including whether it complies with the availability policies of the provider's jurisdiction and your own) are your responsibility. CiCy does not steer you toward, proxy, or embed any channel dedicated to circumventing regional restrictions.\n\n2.3 You must comply with the user agreement, usage policies, and acceptable-use scope of every Third-Party Agent and model service you connect.\n\n---\n\n## 3. Authorization Notice (permissions the Software requires from you)\n\n> To let AI agents do work on your behalf, the Software needs the following authorizations on your device. We group them by risk level and explain each one's **purpose, scope, risk, and whether it can be revoked**. **Transparent disclosure is at the heart of these Terms.**\n\n### 3A. Basic authorizations (granted by accepting these Terms; required for the Software to run)\n\n| Authorization | Purpose | Scope & limits | Revocable |\n|---|---|---|---|\n| **Local file read/write** | Read/write the workspace, configuration, task history | Limited to the Software's working directory and files you explicitly target in a task | Uninstall to end |\n| **Run Third-Party Agents** | Launch sub-processes such as claude/codex/gemini on your computer | Only the Agents you have connected | Remove the Agent |\n| **Local data storage** | Store audit/history/configuration locally | Data stays on your machine; not uploaded | Clear in settings |\n| **Read system & window info** | Scheduling and status display | OS/hardware summary, window list; no file contents | Uninstall to end |\n| **Local network access** | Make requests to the model backend/service you chose, on your behalf | Outbound per your configuration; no embedded circumvention channel | Disable networking in settings |\n\n### 3B. High-risk authorizations (off by default; enabled only after your **separate, explicit** consent)\n\n> Accepting these Terms as a whole will **not** automatically enable any of the following. Each requires a separate action in the app to turn on, and can be turned off at any time.\n\n**① Local command and script execution**\n- **Purpose**: Allow AI agents to execute shell / Python / Node scripts on your computer to complete automation tasks.\n- **Risk (please read carefully)**: This is equivalent to granting a program the **ability to run code** on your computer. Enable it only in tasks you trust.\n- **Controls**: Acts on the **local machine** only; all executions are written to a local audit trail; high-risk commands can require secondary confirmation.\n- **Revoke**: Disable execution in settings.\n\n**② Remote orchestration pairing**\n- **Purpose**: Allow an orchestrator you have **explicitly paired** (which may run in the cloud or on another device) to drive the local Agents over a secure channel, enabling remote collaborative scheduling.\n- **Risk**: Once paired, the remote orchestrator can trigger the local capabilities above.\n- **Controls**: Connects **only to orchestrators you have explicitly paired, with a token you can revoke at any time**; never installs silently, never self-propagates; you can unpair to disconnect immediately.\n- **Revoke**: Unpair / revoke the token.\n\n**③ System root certificate trust store write (requires administrator)**\n- **Purpose**: Required for the HTTPS traffic auditing in Section 4 — installing a **machine-specific** audit certificate into the system trust store.\n- **Risk**: Writing to the system trust root is a highly sensitive operation, so it requires administrator authorization, and the Software **never** pre-installs or distributes a shared certificate (see Section 4).\n- **Revoke**: `uninstall-ca` removes it and clears the consent flag in one step.\n\n**④ Local HTTPS traffic auditing** — see the dedicated Section 4.\n\n### 3C. Authorizations used on demand during tasks (each can be disabled in settings)\n\n| Authorization | Purpose | Scope & limits |\n|---|---|---|\n| **Browser control** | Open/navigate pages, read pages or inject scripts in browser sessions you authorize, read/write cookies | Only browser sessions managed by the Software or authorized by you |\n| **Clipboard read/write** | Pass text within automation flows | As needed by the task; can be disabled |\n| **Screen/window capture** | Let an Agent \"see\" the interface to complete an operation | Screenshots stay local; not uploaded |\n| **Multi-account isolation + per-account proxy** | Isolate cookies/storage/egress IP per identity (e.g. multi-store operations, multi-client isolated delivery) | **For lawful multi-account management only; must not be used for unlawful account farming, traffic faking, or defeating platform risk controls** (see Section 5) |\n\n---\n\n## 4. Local HTTPS Traffic Auditing (optional · off by default · requires explicit consent · requires administrator)\n\n4.1 The Software offers an optional \"Local HTTPS Traffic Auditing\" feature, used to perform local security auditing, sensitive-data redaction, and policy governance on encrypted requests sent from **your machine** to AI providers (such as Claude / OpenAI / DeepSeek / Gemini).\n\n4.2 **This feature is off by default.** It takes effect only when you actively click to enable it in the app and complete administrator authorization. When enabled:\n- The Software generates a **machine-specific, unique** audit certificate on your machine (never a pre-installed or cross-device shared certificate) and requests to install it into your system trust store;\n- Once installed, HTTPS traffic from your machine to the **AI provider domains** above is decrypted locally for auditing; **the scope of decryption is strictly limited to those AI provider domains; all other traffic (banking, social, other websites, etc.) is never decrypted or read**;\n- **Decrypted data is stored only on your machine** and is not uploaded to any server.\n\n4.3 **Revocable at any time**: You can disable the feature and uninstall the certificate (`uninstall-ca`) in settings; after revocation no traffic is decrypted, and the consent flag is cleared.\n\n4.4 **Disclosure obligation in multi-user/team settings**: If you deploy the Software on a device or environment used by multiple people and enable auditing on others' traffic, then you (the deployer) must, on your own and in advance, clearly inform the affected persons and obtain their consent, limited to what is necessary for business management. This disclosure-and-consent obligation rests with you.\n\n## 5. Lawful Use of Automation and Device Control (User Responsibility)\n\n5.1 The capabilities described in Section 3 are general-purpose automation/RPA tool capabilities.\n\n5.2 **Authorization boundary**: These capabilities may act only on devices and accounts **you own or are expressly authorized to use**. Accessing or controlling computer systems you are not authorized for is **strictly prohibited**.\n\n5.3 **Prohibited uses**, including but not limited to:\n- (a) intruding into, unlawfully controlling, or interfering with others' computer information systems;\n- (b) bulk registration, traffic/order faking, \"wool-pulling,\" and other unlawful multi-accounting in violation of platform rules;\n- (c) circumventing or defeating a target platform's risk controls or technical protection measures;\n- (d) data scraping in violation of a target site's robots protocol, terms of use, or applicable law;\n- (e) collecting or processing others' personal information without a lawful basis;\n- (f) any act that violates the laws and regulations of your jurisdiction or that of the target.\n\n5.4 **You bear full responsibility for all operations initiated through the Software and their consequences.** CiCy provides only general-purpose tool capabilities and is not responsible for your specific manner of use; if your use leads to third-party claims or administrative/criminal liability, you shall bear it and hold CiCy harmless.\n\n## 6. Personal Information and Privacy\n\n6.1 The Software processes data on a local-first, minimum-necessary basis and does not collect or upload your work content for operational purposes.\n\n6.2 When you use the Software in scenarios involving personal information (auditing, scraping, or processing traffic that contains personal information), under applicable laws such as the Personal Information Protection Law (China) and the GDPR (EU), **you or your organization may constitute a personal-information processor / data controller** and must fulfill the corresponding obligations on your own (notice and consent, records of processing, responding to data-subject rights, and where necessary data protection impact assessments and breach notifications).\n\n6.3 Where EU personal data is transferred out of the EU, a lawful transfer mechanism required by Chapter V of the GDPR (such as Standard Contractual Clauses) and your explicit consent are required.\n\n6.4 **Region-based backend selection**: To comply with the laws of different jurisdictions, the Software selects a default model backend based on your region (for example, users in mainland China default to model services filed/registered within China); cross-border transfer of personal data requires your separate, explicit consent.\n\n## 7. AI-Generated Content\n\n7.1 Content you generate using the Software and Third-Party Agents is subject to the AI-generated-content labeling obligations of your jurisdiction (such as China's Measures for Labeling AI-Generated Synthetic Content and the transparency requirements of the EU AI Act). CiCy provides labeling/disclosure capabilities where applicable, but the responsibility for compliant use rests with you.\n\n## 8. Compliance and Applicable Jurisdictions\n\n8.1 You must comply with all applicable laws, regulations, and platform terms of your jurisdiction and of the jurisdiction your operations target, **including in particular those of China and the EU**.\n\n8.2 The Software applies different defaults and restrictions to different regions on a \"jurisdiction-aware\" basis, but this does not relieve you of your own responsibility to ensure compliance.\n\n## 9. Disclaimer and Limitation of Liability\n\n9.1 The Software is provided \"as is\"; to the maximum extent permitted by applicable law, no implied warranties of merchantability or fitness for a particular purpose are made.\n\n9.2 To the maximum extent permitted by applicable law, CiCy is not liable for any direct or indirect loss, third-party claim, or administrative or criminal liability arising from your breach of these Terms or from your use of Third-Party Agents / automation capabilities / the auditing feature, and you shall hold CiCy harmless against related claims.\n\n9.3 Nothing in these Terms excludes or limits liability that cannot be excluded under mandatory law (such as liability for personal injury or fraud).\n\n## 10. Consent and Effect\n\n10.1 **These Terms are a precondition to using the Software.** Clicking \"Agree and continue\" means you have read, understood, and accepted all of these Terms; if you click \"Decline and exit\" or make no choice, you **cannot use the Software**.\n\n10.2 The high-risk authorizations in Section 3B (local command/script execution, remote orchestration pairing, root certificate write) and the HTTPS auditing in Section 4 **must be enabled by you through separate actions in the app**; your overall acceptance of these Terms does **not** constitute enabling those features.\n\n10.3 You may review, disable, or revoke each authorization listed in Section 3 at any time in settings.\n\n---\n\n> **Implementation notes (not part of the Terms; for product/legal):**\n> - First launch = blocking terms: show the 6 \"at a glance\" points + \"Agree and continue / Decline and exit\"; no entry to the main UI without acceptance. Full Terms expandable/linked.\n> - **Second, separate consent**: Section 4 auditing = the three-state consent card (not-granted / granted / processing); the Section 3B① execution and 3B② remote-pairing capabilities should likewise have their own switches and first-use confirmation.\n> - Backend gating: Section 4 corresponds to the `~/cicy-ai/mitm/ca-trust-consent` flag + `exec cicy-code mitm install-ca/uninstall-ca`.\n> - Container/headless deployment: if using the `CICY_CA_TRUST_CONSENT=1` escape hatch, \"setting this environment variable means the deployer has, on behalf of its environment, given the notice and consent under Section 4.4\"; responsibility rests with the deployer and this must be documented.\n> - Production: zh + en (EU) bilingual; wording of Sections 3C/6/8 may be adjusted per jurisdiction; high-risk authorizations should retain a user-revocable audit log as evidence of consent.\n",
+};