@foxden-app/foxclaw 0.2.5 → 0.2.7

package/README.md

@@ -1,23 +1,37 @@
 中文 ｜ [English](./README_EN.md)
 
-# FoxClaw
+# 🦊 FoxClaw · 狸爪
 
-FoxClaw 是 Foxden agents 的本地执行爪——跑在你自己电脑上，让你通过 Telegram 或微信远程操控本机的 Codex。
+**狸爪：一个更适配编程需求的移动端 Codex 控制器。**
 
-不需要公网服务器。FoxClaw 在本地和 `codex app-server` 通信，审批在你电脑上完成，结果推送回手机。
+FoxClaw（狸爪）的目标很直接：让你用手机控制本机的 Codex，把移动端变成一个真正能用的 web coding 入口。Telegram 或微信负责交互，`codex app-server` 负责本地执行；你发任务、看进度、批审批、切线程，Codex 在电脑上继续写代码。
 
-## 从这里开始
+这适合你离开办公桌去吃饭、在旅途中、在跑步机上，或者陪小孩逛公园的时候继续工作。人可以离开电脑，Codex 不必停；它会把关键进展、错误、审批请求和最终结果同步回手机。
 
-- 手头有 Codex、OpenClaw、QwenPaw、Hermes、OpenCode、Kimi CLI 之类能跑 shell 的 agent？推荐走 [Agent 辅助安装](./docs/agent-assisted-install.md)。
-- 对 Node、Telegram 机器人、Codex CLI 不太熟？看 [新手安装指南](./docs/install-for-beginners.md)。
-- Git、Node、`.env` 都玩得转？直接往下看快速设置。
-- 卡住了？看 [故障排查](./docs/troubleshooting.md)。
+不需要公网服务器。FoxClaw 跑在你自己的电脑上，代码、shell、认证、审批和运行数据都留在本机，只把你允许的操作入口暴露给受信任的聊天账号。
+
+## 为什么是 FoxClaw
+
+**为什么选 Codex 作为底层引擎？**
+
+1. **开源且接口完整** — Codex 是 OpenAI 开源的 CLI agent，并提供 `codex app-server`。FoxClaw 不是模拟终端输入，而是通过更完整的接口读取线程、切换模型、处理审批和恢复会话。
+2. **当前编码体验强** — FoxClaw 直接读取你本机 Codex 可用的模型列表；如果你的 Codex 环境已经有 GPT-5.5，就可以在手机端选择和使用它。对高强度编程工作流来说，Codex/GPT-5.5 的体感已经是很多人优先选择它的原因。
+3. **多账号额度切换** — 免费额度、一个月 Plus/Team 试用、多账号小额度都可以纳入本地 `auth.json_*` 候选。5 小时限制一到，FoxClaw 自动切换下一个可用账号、重启 app-server 并重试刚才的请求。
 
-FoxClaw 适合这些场景：
+**适合这些场景：**
 
-- 在手机上用 Codex，但不想把电脑暴露到公网
-- 代码、shell、认证、审批、运行数据全部留在本机
-- 只允许一个受信任的 Telegram 用户远程操作
+- 🍜 离开办公桌去吃饭，Codex 继续编码，有审批需求时手机上点一下
+- 🚶 通勤路上、旅途中，不用打开电脑也能下发任务、查看进展、继续调试
+- 🏃 在跑步机上、陪小孩逛公园，持续监控 Codex 的编码进程
+- 🔒 代码、shell、认证、审批、运行数据全部留在本机，不暴露到公网
+- 👤 只允许一个受信任的 Telegram 用户远程操作
+
+## 从这里开始
+
+- 手头有 Codex、OpenClaw、QwenPaw、Hermes、OpenCode、Kimi CLI 之类能跑 shell 的 agent？推荐走 [Agent 辅助安装](./docs/zh/agent-assisted-install.md)。
+- 对 Node、Telegram 机器人、Codex CLI 不太熟？看 [新手安装指南](./docs/zh/install-for-beginners.md)。
+- Git、Node、`.env` 都玩得转？直接往下看快速设置。
+- 卡住了？看 [故障排查](./docs/zh/troubleshooting.md)。
 
 最低要求：一个 Telegram bot token、你的 Telegram 数字用户 ID、Node.js 24、一份已登录的 `codex` CLI。首次安装大约 10–20 分钟。
 
@@ -70,21 +84,45 @@ FoxClaw 只响应 `TG_ALLOWED_USER_ID` 的消息——把机器人拉进群不
 <details>
 <summary>FoxClaw 能做什么</summary>
 
-- 单个 Telegram 用户通过私聊、群组、话题控制
+**核心能力：**
+- 通过 Telegram 私聊、群组、话题控制本地 Codex
 - 可选微信/iLink 通道，复用同一套桥接核心
-- `/threads`、`/open`、`/new`、`/where`、`/interrupt`——稳定的聊天-线程绑定
-- 手机上管理线程生命周期：重命名、归档、取消归档、fork、回滚、compact、review、diff
-- 每个聊天独立的设置面板：模型、reasoning effort、Fast tier、access preset、Agent/Plan 模式、active-turn 行为
-- Codex 账户管理：`/account`、`/quota`、`/login_device`、`/login_cancel`、`/auth add <name>`、`/logout confirm`
-- 触发用量限制时自动在本地 `auth.json_*` 之间切换认证
-- 命令、文件变更、细粒度权限审批的内联按钮
+- 手机上完整管理 Codex 线程生命周期：创建、重命名、归档、fork、回滚、compact、review、diff
+- 命令、文件变更、细粒度权限审批的内联按钮——手机上一键审批
 - MCP elicitation 卡片——工具在 turn 中提出结构化问题时展示
+
+**多账号管理：**
+- Codex 账户管理：`/account`、`/quota`、`/login_device`、`/auth add <name>`
+- 触发用量限制时自动在本地 `auth.json_*` 之间切换认证——5 小时限制到了自动换号
+- `/auth` 面板查看、启用、禁用、切换候选账号
+
+**线程与会话：**
+- `/threads`、`/open`、`/new`、`/where`、`/interrupt`——稳定的聊天-线程绑定
+- 每个聊天独立的设置面板：模型、reasoning effort、Fast tier、access preset、Agent/Plan 模式
 - Skills、MCP、hooks、plugins、apps、feature flags、config、requirements、provider diagnostics
+
+**可靠性：**
 - SQLite 持久化：绑定、offset、审批、待处理提示、审计日志
 - 单实例进程锁，防止同一 bot token 重复 polling
 
 </details>
 
+## 多账号切换
+
+FoxClaw 的一大特色是自动多账号切换。当一个账号的 5 小时用量限制触发时，FoxClaw 会自动切换到下一个可用账号继续工作。
+
+设置方式：
+
+1. 在 Codex 的 auth 目录（通常是 `~/.codex/`）中放置多个认证文件，命名为 `auth.json_personal`、`auth.json_team` 等。
+2. 用 `/auth add <name>` 通过 Telegram 直接添加新账号。
+3. 用 `/auth` 查看所有候选账号状态。
+4. 用 `/auth enable <n>` / `/auth disable <n>` 控制哪些账号参与自动轮换。
+
+当 Codex 报告用量限制错误时，FoxClaw 会自动：
+- 切换到下一个未失败的候选账号
+- 重启 app-server 加载新认证
+- 用新账号重试刚才失败的请求
+
 ## 服务与调试
 
 推荐方式：
@@ -241,11 +279,11 @@ foxclaw weixin-login
 
 ## Codex Skill
 
-仓库自带一个 Codex skill：[`skills/foxclaw`](./skills/foxclaw)。用它可以让 Codex 通过 SSH 在本机或远程 Mac 上 bootstrap FoxClaw——写 `.env`、构建、跑 doctor、装 launchd、引导首次消息验证，一条龙。
+仓库自带一个 Codex skill。用法看 [FoxClaw Skill 中文说明](./docs/zh/foxclaw-skill.md)。它可以让 Codex 通过 SSH 在本机或远程 Mac 上 bootstrap FoxClaw——写 `.env`、构建、跑 doctor、装 launchd、引导首次消息验证，一条龙。
 
 ## 故障排查
 
-`doctor` 报错、Telegram 没回复、服务日志看不懂、重启行为异常、迁移出问题——都看 [故障排查](./docs/troubleshooting.md)。
+`doctor` 报错、Telegram 没回复、服务日志看不懂、重启行为异常、迁移出问题——都看 [故障排查](./docs/zh/troubleshooting.md)。
 
 ## 运维命令
 

package/README_EN.md

@@ -1,10 +1,30 @@
 [中文](./README.md) ｜ English
 
-# FoxClaw
+# 🦊 FoxClaw
 
-FoxClaw is the local execution claw for Foxden agents.
+**A mobile Codex controller built for real programming workflows.**
 
-It runs on your own computer and lets a trusted Telegram or Weixin chat control your local Codex environment. You do not need a public server: FoxClaw talks to Codex over local `codex app-server`, keeps approvals on your machine, and sends the working conversation back to your phone.
+FoxClaw turns your phone into a practical web coding cockpit for your local Codex. Telegram or Weixin handles the chat interface, `codex app-server` handles local execution, and you can send tasks, inspect progress, approve actions, switch threads, and keep working without opening a laptop.
+
+It is built for the moments when you leave your desk for lunch, commute, travel, use a treadmill, or take the kids to the park. You can step away from the keyboard while Codex keeps coding and sends progress, errors, approval requests, and final results back to your phone.
+
+No public server required. FoxClaw runs on your own computer, talks to `codex app-server` locally, and keeps code, shell access, auth, approvals, and runtime data on that machine.
+
+## Why FoxClaw
+
+**Why Codex as the underlying engine?**
+
+1. **Open source with complete APIs** — Codex is OpenAI's open-source CLI agent and ships `codex app-server`. FoxClaw does not scrape a terminal; it uses the app-server interface to read threads, switch models, handle approvals, and resume sessions.
+2. **Strong current coding experience** — FoxClaw reads the model list from your local Codex app-server. If your Codex environment has GPT-5.5 available, you can select and use it from your phone. For many heavy coding workflows, Codex/GPT-5.5 is already the reason to choose this stack.
+3. **Multi-account quota rotation** — Free quota, trial Plus/Team accounts, and small account-specific allowances can all live as local `auth.json_*` candidates. When one account hits a 5-hour usage limit, FoxClaw switches to the next available account, restarts app-server, and retries the failed request.
+
+**Built for these scenarios:**
+
+- 🍜 Leave your desk for lunch — Codex keeps coding, tap to approve when needed
+- 🚶 Commute or travel — dispatch tasks, inspect progress, and continue debugging without a laptop
+- 🏃 Use a treadmill or spend time at the park — monitor Codex's coding process from your phone
+- 🔒 Code, shell, auth, approvals, and runtime data stay on your machine — nothing exposed to the public internet
+- 👤 Only one trusted Telegram user can operate the bot
 
 ## Start Here
 
@@ -13,15 +33,9 @@ It runs on your own computer and lets a trusted Telegram or Weixin chat control
 - Already comfortable with Git, Node, and `.env` files? Use the quick setup below.
 - Something failed? Check [Troubleshooting](./docs/troubleshooting.md).
 
-FoxClaw is a good fit if you want to:
-
-- use Codex from your phone without exposing your computer to the public internet
-- keep code, shell access, auth, approvals, and runtime data on your own machine
-- use one trusted Telegram user as the remote operator
-
-The minimum install needs only a Telegram bot token, your numeric Telegram user id, Node.js 24, and a logged-in `codex` CLI. A first install usually takes 10-20 minutes.
+The minimum install needs only a Telegram bot token, your numeric Telegram user id, Node.js 24, and a logged-in `codex` CLI. A first install usually takes 10–20 minutes.
 
-30-second product example: after FoxClaw is running, send `List files in DEFAULT_CWD` to your Telegram bot. FoxClaw asks local Codex to inspect that folder on your computer and sends the answer back to Telegram.
+**30-second demo**: after FoxClaw is running, send `List files in DEFAULT_CWD` to your Telegram bot. FoxClaw asks local Codex to inspect that folder on your computer and sends the answer back to Telegram.
 
 ## Requirements
 
@@ -31,7 +45,7 @@ The minimum install needs only a Telegram bot token, your numeric Telegram user
 - A Telegram bot token from `@BotFather`
 - Your Telegram numeric user id
 
-## npm Quick Setup
+## Quick Setup
 
 ```bash
 npm install -g @foxden-app/foxclaw
@@ -41,7 +55,7 @@ foxclaw doctor
 foxclaw start
 ```
 
-pnpm users can use:
+pnpm users:
 
 ```bash
 pnpm add -g @foxden-app/foxclaw
@@ -63,28 +77,52 @@ DEFAULT_SANDBOX_MODE=workspace-write
 
 The default config file is `~/.foxclaw/.env`. Set `FOXCLAW_ENV=/path/to/.env` if you want to keep it somewhere else.
 
-`foxclaw start` runs checks and installs or restarts the background service. It is idempotent, so run it again after upgrading.
+`foxclaw start` runs checks and installs or restarts the background service. It is idempotent — run it again after upgrading.
 
 FoxClaw accepts messages only from `TG_ALLOWED_USER_ID`. Putting the bot in a group does not make it available to every group member.
 
 <details>
 <summary>What FoxClaw can do after it is running</summary>
 
-- Telegram private chat, group, and topic control for one allowed Telegram user
-- Optional Weixin/iLink channel for the same bridge core
-- Sticky chat-to-thread binding with `/threads`, `/open`, `/new`, `/where`, and `/interrupt`
-- Thread lifecycle controls from mobile: rename, archive, unarchive, fork, rollback, compact, review, and diff
-- Chat-scoped setup panel for model, reasoning effort, Fast service tier, access preset, Agent/Plan mode, and active-turn behavior
-- Codex account controls with `/account`, `/quota`, `/login_device`, `/login_cancel`, `/auth add <name>`, and guarded `/logout confirm`
-- Automatic local Codex auth rotation across `auth.json_*` candidates when a usage-limit auth fails
-- Inline approval buttons for command, file-change, and granular permission approvals
+**Core capabilities:**
+- Telegram private chat, group, and topic control for your local Codex
+- Optional Weixin/iLink channel sharing the same bridge core
+- Full thread lifecycle management from mobile: create, rename, archive, fork, rollback, compact, review, diff
+- Inline approval buttons for commands, file changes, and granular permissions — one tap to approve
 - MCP elicitation cards for structured questions raised by tools during a turn
+
+**Multi-account management:**
+- Codex account controls: `/account`, `/quota`, `/login_device`, `/auth add <name>`
+- Automatic auth rotation across local `auth.json_*` files when a usage limit is hit — seamless account switching
+- `/auth` panel to view, enable, disable, and switch between candidate accounts
+
+**Threads and sessions:**
+- `/threads`, `/open`, `/new`, `/where`, `/interrupt` — sticky chat-to-thread binding
+- Chat-scoped setup panel for model, reasoning effort, Fast tier, access preset, Agent/Plan mode
 - Skills, MCP, hooks, plugins, apps, feature flags, config, requirements, and provider diagnostics
+
+**Reliability:**
 - SQLite persistence for bindings, offsets, approvals, pending input prompts, and audit logs
 - Single-instance process lock to prevent duplicate Telegram polling on the same bot token
 
 </details>
 
+## Multi-Account Rotation
+
+A key FoxClaw feature is automatic multi-account switching. When one account's 5-hour usage limit is triggered, FoxClaw automatically switches to the next available account and continues working.
+
+Setup:
+
+1. Place multiple auth files in the Codex auth directory (usually `~/.codex/`), named like `auth.json_personal`, `auth.json_team`, etc.
+2. Use `/auth add <name>` to add new accounts directly from Telegram.
+3. Use `/auth` to view all candidate account statuses.
+4. Use `/auth enable <n>` / `/auth disable <n>` to control which accounts participate in auto-rotation.
+
+When Codex reports a usage-limit error, FoxClaw automatically:
+- Switches to the next non-failed candidate account
+- Restarts app-server to load the new auth
+- Retries the failed request with the new account
+
 ## Service And Debugging
 
 Recommended:
@@ -108,13 +146,15 @@ foxclaw serve
 
 Default runtime files are stored under `~/.foxclaw`:
 
-- store: `~/.foxclaw/data/bridge.sqlite`
-- bridge log: `~/.foxclaw/logs/service.log`
-- status: `~/.foxclaw/runtime/status.json`
-- app-server state: `~/.foxclaw/runtime/codex-app-server.json`
-- app-server log: `~/.foxclaw/logs/codex-app-server.log`
+| Purpose | Path |
+|---------|------|
+| Database | `~/.foxclaw/data/bridge.sqlite` |
+| Bridge log | `~/.foxclaw/logs/service.log` |
+| Status | `~/.foxclaw/runtime/status.json` |
+| App-server state | `~/.foxclaw/runtime/codex-app-server.json` |
+| App-server log | `~/.foxclaw/logs/codex-app-server.log` |
 
-Override the store, lock, and app-server paths with `STORE_PATH`, `LOCK_PATH`, `CODEX_APP_SERVER_STATE_PATH`, and `CODEX_APP_SERVER_LOG_PATH`.
+Override with `STORE_PATH`, `LOCK_PATH`, `CODEX_APP_SERVER_STATE_PATH`, and `CODEX_APP_SERVER_LOG_PATH`.
 
 ## Migrating From telegram-codex-app-bridge
 
@@ -156,7 +196,7 @@ TG_ALLOWED_TOPIC_ID=42
 - Set both `TG_ALLOWED_CHAT_ID` and `TG_ALLOWED_TOPIC_ID` to bind one topic as the default scope.
 - Private chat remains available for `TG_ALLOWED_USER_ID` even when a group is configured.
 
-To discover group and topic ids:
+**How to find group and topic IDs:**
 
 1. Stop FoxClaw.
 2. Send a message in the target group or topic.
@@ -197,7 +237,7 @@ No static Codex app-server port is required in normal installs.
 ## Commands
 
 - `/help`
-- `/setup` opens the unified preference panel
+- `/setup` — unified preference panel
 - `/fast <on|off|toggle>`
 - `/active <steer|queue>`
 - `/status`, `/account`, `/quota`
@@ -215,7 +255,7 @@ No static Codex app-server port is required in normal installs.
 - `/skills [query]`, `/skill <name>`, `/skill_enable <name>`, `/skill_disable <name>`
 - `/loaded`, `/hooks`, `/plugins [query]`, `/apps [reload]`, `/features`, `/config`, `/requirements`, `/provider`
 - `/mcp`, `/mcp_reload`, `/mcp_login <server>`, `/mcp_resource <server> <uri>`
-- `/models`, `/model`, `/effort`, `/permissions`, `/access`, `/mode`, `/plan`, and `/agent`
+- `/models`, `/model`, `/effort`, `/permissions`, `/access`, `/mode`, `/plan`, `/agent`
 - `/reveal`, `/where`, `/interrupt`
 
 Plain text sends to the current thread, or creates a new one if none is bound.
@@ -239,7 +279,7 @@ Weixin runtime files default to `~/.foxclaw/weixin`.
 
 ## Codex Skill
 
-This repo ships a Codex skill at [`skills/foxclaw`](./skills/foxclaw). Use it when you want Codex to bootstrap FoxClaw locally or on another Mac over SSH, write `.env`, build, run doctor, install launchd, and guide first-message validation.
+This repo ships a Codex skill at [`skills/foxclaw`](./skills/foxclaw). Use it when you want Codex to bootstrap FoxClaw locally or on another Mac over SSH — write `.env`, build, run doctor, install launchd, and guide first-message validation.
 
 ## Troubleshooting
 
@@ -256,4 +296,4 @@ foxclaw uninstall-systemd
 
 ## Contributing
 
-Issues and PRs are welcome at `https://github.com/foxden-app/foxclaw`.
+Issues and PRs are welcome at [GitHub](https://github.com/foxden-app/foxclaw).

package/dist/controller/controller.js

@@ -6980,7 +6980,7 @@ function renderAuthListMessage(locale, state, includeWeixinCopyPaste = false) {
 function authChoiceKeyboard(locale, record) {
     const rows = record.candidates.map((candidate, index) => [
         {
-            text: clipButtonText(`${candidate.isCurrent ? '* ' : ''}${candidate.name}${candidate.disabled ? ' off' : ''}`),
+            text: clipButtonText(`${candidate.isCurrent ? '✅ ' : '🔐 '}${candidate.name}${candidate.disabled ? ' · off' : ''}`),
             callback_data: `auth:${record.localId}:${index}`,
         },
         {

package/dist/controller/presentation.js

@@ -38,7 +38,7 @@ export function buildThreadsKeyboard(locale, threads) {
     return threads.flatMap((thread, index) => {
         const ordinal = typeof thread.index === 'number' ? thread.index : index + 1;
         const openRow = [{
-                text: `${ordinal}. ${truncate(compactWhitespace(thread.name || thread.preview || t(locale, 'empty')), 28)}`,
+                text: `🧵 ${ordinal}. ${truncate(compactWhitespace(thread.name || thread.preview || t(locale, 'empty')), 28)}`,
                 callback_data: `thread:open:${thread.threadId}`,
             }];
         const actionRow = thread.archived
@@ -110,15 +110,15 @@ export function buildAccessSettingsKeyboard(locale, access) {
     const currentPreset = access.preset;
     const buttons = [
         {
-            text: `${currentPreset === 'read-only' ? '• ' : ''}${t(locale, 'access_preset_read_only')}`,
+            text: selectedButtonText(currentPreset === 'read-only', accessPresetButtonLabel(locale, 'read-only')),
             callback_data: 'settings:access:read-only',
         },
         {
-            text: `${currentPreset === 'default' ? '• ' : ''}${t(locale, 'access_preset_default')}`,
+            text: selectedButtonText(currentPreset === 'default', accessPresetButtonLabel(locale, 'default')),
             callback_data: 'settings:access:default',
         },
         {
-            text: `${currentPreset === 'full-access' ? '• ' : ''}${t(locale, 'access_preset_full_access')}`,
+            text: selectedButtonText(currentPreset === 'full-access', accessPresetButtonLabel(locale, 'full-access')),
             callback_data: 'settings:access:full-access',
         },
     ];
@@ -256,7 +256,7 @@ export function buildSetupPanelKeyboard(locale, ctx) {
             callback_data: 'setup:model:default',
         },
         ...ctx.models.map((model) => ({
-            text: `${currentModel === model.model ? '• ' : ''}${truncate(model.model, 14)}`,
+            text: selectedButtonText(currentModel === model.model, `🤖 ${truncate(model.model, 14)}`),
             callback_data: `setup:model:${encodeURIComponent(model.model)}`,
         })),
     ], 2));
@@ -266,7 +266,7 @@ export function buildSetupPanelKeyboard(locale, ctx) {
             callback_data: 'setup:effort:default',
         },
         ...efforts.map((effort) => ({
-            text: `${ctx.settings?.reasoningEffort === effort ? '• ' : ''}${effort}`,
+            text: selectedButtonText(ctx.settings?.reasoningEffort === effort, `🧠 ${effort}`),
             callback_data: `setup:effort:${effort}`,
         })),
     ], 3));
@@ -287,35 +287,35 @@ export function buildSetupPanelKeyboard(locale, ctx) {
     }
     rows.push([
         {
-            text: `${ctx.access.preset === 'read-only' ? '• ' : ''}${t(locale, 'access_preset_read_only')}`,
+            text: selectedButtonText(ctx.access.preset === 'read-only', accessPresetButtonLabel(locale, 'read-only')),
             callback_data: 'setup:access:read-only',
         },
         {
-            text: `${ctx.access.preset === 'default' ? '• ' : ''}${t(locale, 'access_preset_default')}`,
+            text: selectedButtonText(ctx.access.preset === 'default', accessPresetButtonLabel(locale, 'default')),
             callback_data: 'setup:access:default',
         },
         {
-            text: `${ctx.access.preset === 'full-access' ? '• ' : ''}${t(locale, 'access_preset_full_access')}`,
+            text: selectedButtonText(ctx.access.preset === 'full-access', accessPresetButtonLabel(locale, 'full-access')),
             callback_data: 'setup:access:full-access',
         },
     ]);
     rows.push([
         {
-            text: `${currentMode === 'default' ? '• ' : ''}${t(locale, 'collaboration_mode_default')}`,
+            text: selectedButtonText(currentMode === 'default', `🤖 ${t(locale, 'collaboration_mode_default')}`),
             callback_data: 'setup:mode:default',
         },
         {
-            text: `${currentMode === 'plan' ? '• ' : ''}${t(locale, 'collaboration_mode_plan')}`,
+            text: selectedButtonText(currentMode === 'plan', `📝 ${t(locale, 'collaboration_mode_plan')}`),
             callback_data: 'setup:mode:plan',
         },
     ]);
     rows.push([
         {
-            text: `${activeTurnMessageMode === 'steer' ? '• ' : ''}${t(locale, 'active_turn_message_mode_steer')}`,
+            text: selectedButtonText(activeTurnMessageMode === 'steer', `🎯 ${t(locale, 'active_turn_message_mode_steer')}`),
             callback_data: 'setup:active:steer',
         },
         {
-            text: `${activeTurnMessageMode === 'queue' ? '• ' : ''}${t(locale, 'active_turn_message_mode_queue')}`,
+            text: selectedButtonText(activeTurnMessageMode === 'queue', `📥 ${t(locale, 'active_turn_message_mode_queue')}`),
             callback_data: 'setup:active:queue',
         },
     ]);
@@ -382,6 +382,16 @@ function formatFastSetupLabel(locale, supported, enabled, tierName) {
 function resolveCollaborationMode(mode) {
     return mode === 'plan' ? 'plan' : 'default';
 }
+function selectedButtonText(selected, label) {
+    return `${selected ? '• ' : ''}${label}`;
+}
+function accessPresetButtonLabel(locale, preset) {
+    if (preset === 'read-only')
+        return `👁️ ${formatAccessPresetLabel(locale, preset)}`;
+    if (preset === 'full-access')
+        return `🔓 ${formatAccessPresetLabel(locale, preset)}`;
+    return `🛡️ ${formatAccessPresetLabel(locale, preset)}`;
+}
 export function formatModelSettingsMessage(locale, models, settings) {
     const selectedModel = resolveCurrentModel(models, settings?.model ?? null);
     const selectedModelLabel = settings?.model ?? t(locale, 'server_default');
@@ -417,7 +427,7 @@ export function buildModelSettingsKeyboard(locale, models, settings) {
             callback_data: 'settings:model:default',
         },
         ...models.map((model) => ({
-            text: `${currentModel === model.model ? '• ' : ''}${truncate(model.model, 14)}`,
+            text: selectedButtonText(currentModel === model.model, `🤖 ${truncate(model.model, 14)}`),
             callback_data: `settings:model:${encodeURIComponent(model.model)}`,
         })),
     ];
@@ -427,7 +437,7 @@ export function buildModelSettingsKeyboard(locale, models, settings) {
             callback_data: 'settings:effort:default',
         },
         ...efforts.map((effort) => ({
-            text: `${settings?.reasoningEffort === effort ? '• ' : ''}${effort}`,
+            text: selectedButtonText(settings?.reasoningEffort === effort, `🧠 ${effort}`),
             callback_data: `settings:effort:${effort}`,
         })),
     ];

package/dist/i18n.d.ts

@@ -148,10 +148,10 @@ declare const MESSAGES: {
         readonly auth_add_cancelled: "New auth login cancelled. Restored previous auth.";
         readonly auth_add_reverted: "Restored previous auth.";
         readonly auth_add_missing_file: "Login completed, but the new auth file was not created: {value}";
-        readonly button_login_device: "Login device";
-        readonly button_auth_reload: "Reload auth";
-        readonly button_auth_enable: "Enable";
-        readonly button_auth_disable: "Disable";
+        readonly button_login_device: "🔑 Login";
+        readonly button_auth_reload: "🔄 Reload auth";
+        readonly button_auth_enable: "✅ Enable";
+        readonly button_auth_disable: "⏸️ Disable";
         readonly another_turn_running: "Another turn is already running. Use /interrupt, /takeover, /queue, or wait.";
         readonly working: "Working...";
         readonly usage_open: "Usage: /open <n>";
@@ -295,16 +295,16 @@ declare const MESSAGES: {
         readonly approval_decision_accept: "allow";
         readonly approval_decision_session: "allow for session";
         readonly approval_decision_deny: "deny";
-        readonly button_allow: "Allow";
-        readonly button_allow_session: "Allow Session";
-        readonly button_deny: "Deny";
-        readonly button_interrupt: "Interrupt";
-        readonly button_permissions: "Access";
-        readonly button_models: "Models";
-        readonly button_threads: "Threads";
-        readonly button_reveal: "Reveal";
-        readonly button_auto: "Auto";
-        readonly button_new_thread: "New";
+        readonly button_allow: "✅ Allow";
+        readonly button_allow_session: "✅ Session";
+        readonly button_deny: "🚫 Deny";
+        readonly button_interrupt: "⏹️ Interrupt";
+        readonly button_permissions: "🛡️ Access";
+        readonly button_models: "🤖 Models";
+        readonly button_threads: "🧵 Threads";
+        readonly button_reveal: "↗️ Reveal";
+        readonly button_auto: "⚙️ Auto";
+        readonly button_new_thread: "➕ New";
         readonly threads_no_matches: "<b>No threads matched</b>\n<code>{searchTerm}</code>";
         readonly threads_no_recent: "<b>No recent threads.</b>";
         readonly threads_recent_title: "<b>Recent threads</b>";
@@ -313,15 +313,15 @@ declare const MESSAGES: {
         readonly threads_filter: "Filter: <code>{searchTerm}</code>";
         readonly threads_range: "Showing {start}-{end}";
         readonly threads_filter_cleared_short: "Filter cleared";
-        readonly button_prev_page: "Prev";
-        readonly button_next_page: "Next";
-        readonly button_clear_filter: "Clear filter";
-        readonly button_recent_threads: "Recent";
-        readonly button_archived_threads: "Archived";
-        readonly button_thread_rename: "Rename";
-        readonly button_thread_watch: "Watch";
-        readonly button_thread_archive: "Archive/Delete";
-        readonly button_thread_unarchive: "Unarchive";
+        readonly button_prev_page: "⬅️ Prev";
+        readonly button_next_page: "➡️ Next";
+        readonly button_clear_filter: "🧹 Clear";
+        readonly button_recent_threads: "🕘 Recent";
+        readonly button_archived_threads: "🗄️ Archived";
+        readonly button_thread_rename: "✏️ Rename";
+        readonly button_thread_watch: "👀 Watch";
+        readonly button_thread_archive: "🗑️ Archive/Delete";
+        readonly button_thread_unarchive: "♻️ Unarchive";
         readonly threads_current: "Current: <b>{title}</b>";
         readonly where_no_thread_bound: "No thread is currently bound.";
         readonly where_send_message_or_new: "Send a message or use /new.";
@@ -370,8 +370,8 @@ declare const MESSAGES: {
         readonly active_turn_message_mode_steer: "Steer current turn";
         readonly active_turn_message_mode_queue: "Queue next turn";
         readonly button_fast_on: "⚡ Fast: on";
-        readonly button_fast_off: "Fast: off";
-        readonly button_fast_unsupported: "Fast unsupported";
+        readonly button_fast_off: "⚪ Fast: off";
+        readonly button_fast_unsupported: "⛔ Fast unsupported";
         readonly access_preset_read_only: "Read-only";
         readonly access_preset_default: "Default";
         readonly access_preset_full_access: "Full access";
@@ -506,10 +506,10 @@ declare const MESSAGES: {
         readonly permission_read_paths: "Read paths: {value}";
         readonly permission_write_paths: "Write paths: {value}";
         readonly permission_entries: "File entries: {value}";
-        readonly button_accept: "Accept";
-        readonly button_decline: "Decline";
-        readonly button_cancel: "Cancel";
-        readonly button_create_dir: "Create";
+        readonly button_accept: "✅ Accept";
+        readonly button_decline: "🚫 Decline";
+        readonly button_cancel: "✖️ Cancel";
+        readonly button_create_dir: "📁 Create";
         readonly thread_name_updated: "Thread renamed: {name}";
         readonly thread_archived_notification: "Thread archived: {threadId}";
         readonly thread_unarchived_notification: "Thread unarchived: {threadId}";
@@ -696,10 +696,10 @@ declare const MESSAGES: {
         readonly auth_add_cancelled: "新 auth 登录已取消，已恢复之前的 auth。";
         readonly auth_add_reverted: "已恢复之前的 auth。";
         readonly auth_add_missing_file: "登录已完成，但没有创建新的 auth 文件：{value}";
-        readonly button_login_device: "设备登录";
-        readonly button_auth_reload: "重载 auth";
-        readonly button_auth_enable: "启用";
-        readonly button_auth_disable: "禁用";
+        readonly button_login_device: "🔑 设备登录";
+        readonly button_auth_reload: "🔄 重载 auth";
+        readonly button_auth_enable: "✅ 启用";
+        readonly button_auth_disable: "⏸️ 禁用";
         readonly another_turn_running: "已经有一个回复在进行中。请先等待，或使用 /interrupt、/takeover、/queue。";
         readonly working: "处理中...";
         readonly usage_open: "用法：/open <编号>";
@@ -843,16 +843,16 @@ declare const MESSAGES: {
         readonly approval_decision_accept: "允许";
         readonly approval_decision_session: "本会话内允许";
         readonly approval_decision_deny: "拒绝";
-        readonly button_allow: "允许";
-        readonly button_allow_session: "本会话允许";
-        readonly button_deny: "拒绝";
-        readonly button_interrupt: "中断";
-        readonly button_permissions: "权限";
-        readonly button_models: "模型";
-        readonly button_threads: "线程";
-        readonly button_reveal: "打开";
-        readonly button_auto: "自动";
-        readonly button_new_thread: "新建";
+        readonly button_allow: "✅ 允许";
+        readonly button_allow_session: "✅ 本会话";
+        readonly button_deny: "🚫 拒绝";
+        readonly button_interrupt: "⏹️ 中断";
+        readonly button_permissions: "🛡️ 权限";
+        readonly button_models: "🤖 模型";
+        readonly button_threads: "🧵 线程";
+        readonly button_reveal: "↗️ 打开";
+        readonly button_auto: "⚙️ 自动";
+        readonly button_new_thread: "➕ 新建";
         readonly threads_no_matches: "<b>没有匹配的线程</b>\n<code>{searchTerm}</code>";
         readonly threads_no_recent: "<b>暂无最近线程。</b>";
         readonly threads_recent_title: "<b>最近线程</b>";
@@ -861,15 +861,15 @@ declare const MESSAGES: {
         readonly threads_filter: "筛选：<code>{searchTerm}</code>";
         readonly threads_range: "显示第 {start}-{end} 条";
         readonly threads_filter_cleared_short: "已清除筛选";
-        readonly button_prev_page: "上一页";
-        readonly button_next_page: "下一页";
-        readonly button_clear_filter: "清除筛选";
-        readonly button_recent_threads: "最近线程";
-        readonly button_archived_threads: "已归档";
-        readonly button_thread_rename: "重命名";
-        readonly button_thread_watch: "观察";
-        readonly button_thread_archive: "归档/删除";
-        readonly button_thread_unarchive: "取消归档";
+        readonly button_prev_page: "⬅️ 上一页";
+        readonly button_next_page: "➡️ 下一页";
+        readonly button_clear_filter: "🧹 清除";
+        readonly button_recent_threads: "🕘 最近线程";
+        readonly button_archived_threads: "🗄️ 已归档";
+        readonly button_thread_rename: "✏️ 重命名";
+        readonly button_thread_watch: "👀 观察";
+        readonly button_thread_archive: "🗑️ 归档/删除";
+        readonly button_thread_unarchive: "♻️ 取消归档";
         readonly threads_current: "当前：<b>{title}</b>";
         readonly where_no_thread_bound: "当前没有绑定线程。";
         readonly where_send_message_or_new: "直接发一条消息，或者使用 /new。";
@@ -918,8 +918,8 @@ declare const MESSAGES: {
         readonly active_turn_message_mode_steer: "引导当前回复";
         readonly active_turn_message_mode_queue: "排队到下一轮";
         readonly button_fast_on: "⚡ Fast：开";
-        readonly button_fast_off: "Fast：关";
-        readonly button_fast_unsupported: "Fast 不支持";
+        readonly button_fast_off: "⚪ Fast：关";
+        readonly button_fast_unsupported: "⛔ Fast 不支持";
         readonly access_preset_read_only: "只读";
         readonly access_preset_default: "默认";
         readonly access_preset_full_access: "完全访问";
@@ -1054,10 +1054,10 @@ declare const MESSAGES: {
         readonly permission_read_paths: "可读路径：{value}";
         readonly permission_write_paths: "可写路径：{value}";
         readonly permission_entries: "文件条目：{value}";
-        readonly button_accept: "接受";
-        readonly button_decline: "拒绝";
-        readonly button_cancel: "取消";
-        readonly button_create_dir: "新建";
+        readonly button_accept: "✅ 接受";
+        readonly button_decline: "🚫 拒绝";
+        readonly button_cancel: "✖️ 取消";
+        readonly button_create_dir: "📁 新建";
         readonly thread_name_updated: "线程已重命名：{name}";
         readonly thread_archived_notification: "线程已归档：{threadId}";
         readonly thread_unarchived_notification: "线程已取消归档：{threadId}";