@foxden-app/foxclaw 0.4.5 → 0.4.7

package/CHANGELOG.md

@@ -0,0 +1,195 @@
+# Changelog
+
+All notable FoxClaw changes are listed here. Each release note is bilingual so GitHub Releases and the npm package are useful to both Chinese and English readers.
+
+## 0.4.7 - 2026-06-02
+
+### 中文
+- `/auth` 面板新增“刷新全部”按钮，并支持 `/auth refresh all`。
+- 刷新全部仅在所有 Telegram runtime、微信 runtime、审批、待输入、登录流程和 auth 镜像写入都空闲时执行。
+- 执行时逐个切换 ChatGPT 候选，调用 Codex `account/read refreshToken=true` 主动刷新，使用 usage 接口验证后再镜像同步，并在结束后恢复原当前 auth、刷新面板摘要。
+- 更新 FoxClaw/npm 发布 skill：用户要求收尾或发布时，校验通过后自动 commit、push、publish，只在真实外部阻塞时再等待用户。
+
+### English
+- Added a Refresh all button to the `/auth` panel and the `/auth refresh all` command.
+- Refresh all runs only when every Telegram runtime, the Weixin runtime, approvals, inputs, login flows, and auth mirror writes are idle.
+- It visits each ChatGPT candidate, calls Codex `account/read refreshToken=true`, validates through the usage endpoint before mirroring, restores the original current auth, and refreshes the panel summary.
+- Updated the FoxClaw/npm publish skills so release wrap-up proceeds through commit, push, and publish after validation, pausing only for real external blockers.
+
+## 0.4.6 - 2026-06-01
+
+### 中文
+- 新增公开 `CHANGELOG.md`，让 GitHub 项目首页和 npm 包都能直接看到版本更新记录。
+- README 增加更新日志入口，并将 `CHANGELOG.md` 纳入 npm 发布包。
+- 发布 workflow 增加 GitHub Release 创建/更新步骤，后续 tag 发布会自动带上对应版本说明。
+
+### English
+- Added a public `CHANGELOG.md` so release notes are visible from GitHub and the npm package.
+- Linked the changelog from the README and included it in the npm package contents.
+- Added GitHub Release creation/update to the publish workflow so future version tags carry release notes automatically.
+
+## 0.4.5 - 2026-06-01
+
+### 中文
+- `/auth` 面板点击切换账号后会刷新原消息并保留按钮，方便连续切换多个候选账号。
+- 回归测试覆盖同一个面板上连续切换的行为。
+
+### English
+- The `/auth` panel now refreshes the original message and keeps its buttons after switching accounts, making consecutive switches practical.
+- Added regression coverage for repeated switches from the same panel.
+
+## 0.4.4 - 2026-06-01
+
+### 中文
+- 多 bot 模式下，切换或重载认证前会从其他 bot home 和规范凭据目录恢复同账号的较新可用候选。
+- 认证同步前增加在线 usage 验证，避免只看 `last_refresh` 时传播已经失效的凭据。
+- 登录/刷新成功后继续将同账号新凭据镜像到其他 runtime，解决某个 bot 仍拿到旧 401 凭据的问题。
+
+### English
+- In multi-bot mode, auth switch/reload now recovers a newer same-account candidate from other bot homes or the canonical auth directory before restart.
+- Added online usage validation before syncing refreshed credentials, avoiding propagation based only on `last_refresh`.
+- Successful login/refresh still mirrors the validated same-account credential to other runtimes, fixing stale 401 credentials in sibling bots.
+
+## 0.4.3 - 2026-05-29
+
+### 中文
+- `/auth` 在多 bot 模式下会按 ChatGPT account id 汇总各 runtime 最近掌握的额度快照。
+- 不同 bot 使用不同 auth 时，面板可以互补展示更完整的 5 小时/7 天额度信息。
+
+### English
+- `/auth` now merges recent quota snapshots by ChatGPT account id across runtimes in multi-bot mode.
+- When different bots use different auth candidates, the panel can present a more complete 5-hour/7-day quota view.
+
+## 0.4.2 - 2026-05-28
+
+### 中文
+- 支持把 `TG_BOT_TOKENS` 中的一个 bot 标记为默认/终端共享 runtime。
+- 当 `TG_BOT_TOKEN` 精确匹配复数 token 中的一个值时，该 bot 使用终端默认 `CODEX_HOME` 和默认 auth，其他 bot 继续隔离。
+- 文档补充共享 runtime 的配置方式和注意事项。
+
+### English
+- Added support for marking one `TG_BOT_TOKENS` bot as the default/shared-terminal runtime.
+- When `TG_BOT_TOKEN` exactly matches one token in the multi-token list, that bot uses the terminal/default `CODEX_HOME` and auth while the others stay isolated.
+- Updated docs with shared-runtime setup guidance.
+
+## 0.4.1 - 2026-05-28
+
+### 中文
+- 认证候选刷新后，镜像同步前会校验同账号和可用性，降低错误凭据扩散风险。
+- 增加认证刷新验证相关测试。
+
+### English
+- Refreshed auth candidates are validated for same-account usability before mirroring, reducing the risk of spreading bad credentials.
+- Added tests for refreshed auth validation.
+
+## 0.4.0 - 2026-05-27
+
+### 中文
+- 完善多 Telegram bot 隔离 runtime 的状态展示、验收路径和文档。
+- 补充多 bot auth 镜像、独立 `CODEX_HOME`、群聊点名规则和全局 `/update` 空闲检查说明。
+
+### English
+- Completed status reporting, validation paths, and docs for isolated multi-Telegram-bot runtimes.
+- Documented multi-bot auth mirroring, independent `CODEX_HOME`, group addressing rules, and global `/update` idle checks.
+
+## 0.3.19 - 2026-05-27
+
+### 中文
+- 支持一个 FoxClaw 服务同时运行多个 Telegram bot，每个 bot 拥有独立 Codex app-server、独立 `CODEX_HOME`、独立会话和独立当前 auth。
+- Telegram scope 加入 bot namespace，避免私聊、群聊、按钮回调和附件处理串线。
+- 新增认证候选镜像协调器，仅同步经过校验的同账号候选文件，不共享 Codex sessions。
+- `/update` 改为全服务空闲检查，并会尝试同步升级 npm/pnpm 管理的 Codex CLI。
+
+### English
+- Added support for running multiple Telegram bots in one FoxClaw service, with independent Codex app-servers, `CODEX_HOME` directories, sessions, and active auth selections.
+- Added bot namespaces to Telegram scopes so private chats, groups, callbacks, and media handling do not cross runtimes.
+- Added an auth candidate mirror coordinator that syncs validated same-account candidates without sharing Codex sessions.
+- Made `/update` a global idle-checked operation and added npm/pnpm-managed Codex CLI update attempts.
+
+## 0.3.18 - 2026-05-27
+
+### 中文
+- 修正 Codex 本地答复吞吐统计口径。
+
+### English
+- Corrected local Codex response throughput metrics.
+
+## 0.3.17 - 2026-05-26
+
+### 中文
+- 修复 GitHub 发布流程的 checkout 鉴权失败。
+
+### English
+- Fixed checkout authentication failure in the GitHub release workflow.
+
+## 0.3.16 - 2026-05-26
+
+### 中文
+- 缓存 `/status` 本地用量统计快照，减少状态页等待时间并优化响应。
+
+### English
+- Cached local usage statistics for `/status`, reducing wait time and improving responsiveness.
+
+## 0.3.15 - 2026-05-26
+
+### 中文
+- 修复 pnpm 全局安装下 `/update` 对 Codex CLI 真实安装路径的识别。
+
+### English
+- Fixed `/update` detection of the real Codex CLI install path under pnpm global installs.
+
+## 0.3.14 - 2026-05-26
+
+### 中文
+- 新增 FoxClaw 一键自升级和聊天内 `/update` 命令。
+
+### English
+- Added one-command FoxClaw self-update and the chat `/update` command.
+
+## 0.3.13 - 2026-05-26
+
+### 中文
+- 新增 Codex 输出速度展示和线程上下文回显。
+
+### English
+- Added Codex output throughput display and thread context echoes.
+
+## 0.3.12 - 2026-05-26
+
+### 中文
+- 简化 FoxClaw 代理服务配置。
+
+### English
+- Simplified FoxClaw proxy service configuration.
+
+## 0.3.11 - 2026-05-26
+
+### 中文
+- 修复 systemd 安装时覆盖旧版本路径的问题。
+
+### English
+- Fixed systemd install behavior that could overwrite an older version path incorrectly.
+
+## 0.3.10 - 2026-05-26
+
+### 中文
+- 支持 npm 发布令牌回退，用于 trusted publishing 不可用时的补发路径。
+
+### English
+- Added npm token fallback support for manual recovery when trusted publishing is unavailable.
+
+## 0.1.1 - 2026-03-06
+
+### 中文
+- 显式解析 Codex 二进制路径，提升运行时定位可靠性。
+
+### English
+- Resolved the Codex binary path explicitly for more reliable runtime startup.
+
+## 0.1.0 - 2026-03-06
+
+### 中文
+- 初始化 Telegram Codex app bridge。
+
+### English
+- Bootstrapped the Telegram Codex app bridge.

package/README.md

@@ -37,6 +37,7 @@ FoxClaw（狸爪）的目标很直接：让你用手机控制本机的 Codex，
 - 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)。
 - 已经装好，想系统了解 `/help`、`/setup`、`/threads`、`/watch`、`/auth` 和账号轮转？看 [用户手册](./docs/zh/user-manual.md)。
+- 想了解每个版本改了什么？看 [更新日志](./CHANGELOG.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)。
 
@@ -257,7 +258,7 @@ FoxClaw 会把 `codex app-server` 作为 detached 子进程启动，记录其 pi
 - `/status`、`/account`、`/quota`、`/update`
 - `/quota_nudge <credits|usage_limit> confirm`
 - `/login_device`、`/login_cancel [id]`、`/logout confirm`
-- `/auth [list|use <n>|enable <n>|disable <n>|reload|add <name>]`
+- `/auth [list|use <n>|enable <n>|disable <n>|reload|refresh all|add <name>]`
 - `/threads [query]`、`/threads archived`、`/open <n>`
 - `/goal [objective|pause|resume|done|budget <tokens|off>|clear confirm]`
 - `/history [limit]`、`/files <query>`、`/remote`

package/README_EN.md

@@ -37,6 +37,7 @@ FoxClaw is more than message forwarding. It provides Telegram panels for Codex w
 - Already have a shell-capable agent such as Codex, OpenClaw, QwenPaw, Hermes, OpenCode, or Kimi CLI? Use the [Agent-Assisted Install](./docs/agent-assisted-install.md) first. This is the recommended path.
 - New to Node, Telegram bots, or Codex CLI? Use the [Beginner Install Guide](./docs/install-for-beginners.md).
 - Already installed and want the full command guide for `/help`, `/setup`, `/threads`, `/watch`, `/auth`, and auth rotation? Read the [User Manual](./docs/user-manual.md).
+- Want to see what changed in each release? Read the [Changelog](./CHANGELOG.md).
 - Already comfortable with Git, Node, and `.env` files? Use the quick setup below.
 - Something failed? Check [Troubleshooting](./docs/troubleshooting.md).
 
@@ -257,7 +258,7 @@ No static Codex app-server port is required in normal installs.
 - `/status`, `/account`, `/quota`, `/update`
 - `/quota_nudge <credits|usage_limit> confirm`
 - `/login_device`, `/login_cancel [id]`, `/logout confirm`
-- `/auth [list|use <n>|enable <n>|disable <n>|reload|add <name>]`
+- `/auth [list|use <n>|enable <n>|disable <n>|reload|refresh all|add <name>]`
 - `/threads [query]`, `/threads archived`, `/open <n>`
 - `/goal [objective|pause|resume|done|budget <tokens|off>|clear confirm]`
 - `/history [limit]`, `/files <query>`, `/remote`

package/dist/codex_app/client.d.ts

@@ -125,7 +125,7 @@ export declare class CodexAppClient extends EventEmitter {
         reviewThreadId: string;
     }>;
     listModels(): Promise<ModelInfo[]>;
-    readAccount(): Promise<CodexAccountInfo | null>;
+    readAccount(refreshToken?: boolean): Promise<CodexAccountInfo | null>;
     startDeviceLogin(): Promise<CodexLoginDeviceCode>;
     cancelLogin(loginId: string): Promise<void>;
     logoutAccount(): Promise<void>;

package/dist/codex_app/client.js

@@ -277,8 +277,8 @@ export class CodexAppClient extends EventEmitter {
         } while (cursor);
         return models;
     }
-    async readAccount() {
-        const result = await this.request('account/read', { refreshToken: false });
+    async readAccount(refreshToken = false) {
+        const result = await this.request('account/read', { refreshToken });
         const account = result?.account;
         if (!account || typeof account !== 'object') {
             return null;

package/dist/controller/controller.d.ts

@@ -45,6 +45,7 @@ export declare class BridgeSessionCore {
     private pendingApprovalMessages;
     private pendingAuthRotation;
     private authRotationInProgress;
+    private authRefreshAllInProgress;
     private authRotationFailedTargets;
     private localUsageCache;
     private localUsageCacheLoaded;
@@ -231,12 +232,14 @@ export declare class BridgeSessionCore {
     private handleFastCommand;
     private setCollaborationMode;
     private handleAuthReloadCommand;
+    private canRunGlobalAuthRefresh;
     private handleSelfUpdateCommand;
     private scheduleSelfUpdateStatusPoll;
     private clearSelfUpdateStatusPoll;
     private pollSelfUpdateStatus;
     private formatSelfUpdateResult;
     private handleAuthCommand;
+    private handleAuthRefreshAllCommand;
     private handleAuthUseCommand;
     private handleAuthToggleCommand;
     private findLatestAuthChoiceList;
@@ -288,6 +291,7 @@ export declare class BridgeSessionCore {
     private switchCodexAuthAndRestart;
     private recoverCodexAuthCandidate;
     private syncCodexAuthCandidate;
+    private refreshAllCodexAuthCandidates;
     private buildNativeCollaborationMode;
     private buildCodexUsageStatusLines;
     private buildCodexLocalUsageStatusLines;
@@ -299,6 +303,7 @@ export declare class BridgeSessionCore {
     private codexLocalUsageSnapshotPath;
     private refreshCurrentCodexAuthQuota;
     private applySharedCodexAuthQuotaSnapshots;
+    private recordCodexAuthQuotaSnapshot;
     private readCodexAuthCandidateAccountIds;
     private codexAuthQuotaSnapshotMatchesAccount;
     private readCodexAuthQuotaSnapshots;

package/dist/controller/controller.js

@@ -113,6 +113,7 @@ export class BridgeSessionCore {
     pendingApprovalMessages = new Map();
     pendingAuthRotation = null;
     authRotationInProgress = false;
+    authRefreshAllInProgress = false;
     authRotationFailedTargets = new Set();
     localUsageCache = null;
     localUsageCacheLoaded = false;
@@ -232,6 +233,7 @@ export class BridgeSessionCore {
         this.recentCommandUsageByScope.clear();
         this.commandUsageSequence = 0;
         this.pendingAuthRotation = null;
+        this.authRefreshAllInProgress = false;
         this.clearObservedThreadWatchers();
         this.releaseActiveTurnsForBridgeShutdown();
         this.bot.stop();
@@ -1008,7 +1010,7 @@ export class BridgeSessionCore {
             await this.handleAuthToggleCallback(event, authToggleMatch[1], Number.parseInt(authToggleMatch[2], 10), locale);
             return;
         }
-        const authActionMatch = /^auth:([a-f0-9]+):(login_device|reload)$/.exec(event.data);
+        const authActionMatch = /^auth:([a-f0-9]+):(login_device|reload|refresh_all)$/.exec(event.data);
         if (authActionMatch) {
             await this.handleAuthPanelActionCallback(event, authActionMatch[1], authActionMatch[2], locale);
             return;
@@ -2604,7 +2606,8 @@ export class BridgeSessionCore {
             && this.pendingUserInputs.size === 0
             && this.pendingMcpElicitations.size === 0
             && this.pendingLoginsByScope.size === 0
-            && !this.authRotationInProgress;
+            && !this.authRotationInProgress
+            && !this.authRefreshAllInProgress;
     }
     hasLocalBlockingActivity() {
         return !this.isIdleForServiceUpdate();
@@ -3779,6 +3782,12 @@ export class BridgeSessionCore {
         lines.push(...await this.buildCodexUsageStatusLines(locale));
         await this.sendMessage(scopeId, lines.join('\n'));
     }
+    canRunGlobalAuthRefresh() {
+        return this.isIdleForServiceUpdate()
+            && this.store.countPendingApprovals() === 0
+            && this.store.countPendingUserInputs() === 0
+            && (!this.coordinator?.canSelfUpdate || this.coordinator.canSelfUpdate());
+    }
     async handleSelfUpdateCommand(scopeId, locale) {
         if (!this.selfUpdater) {
             await this.sendMessage(scopeId, t(locale, 'update_unavailable'));
@@ -3860,6 +3869,18 @@ export class BridgeSessionCore {
             await this.handleAuthReloadCommand(scopeId, locale);
             return;
         }
+        if (action === 'refresh') {
+            if (args[1]?.toLowerCase() === 'all') {
+                await this.handleAuthRefreshAllCommand(scopeId, locale);
+                return;
+            }
+            await this.sendMessage(scopeId, t(locale, 'usage_auth'));
+            return;
+        }
+        if (action === 'refresh_all') {
+            await this.handleAuthRefreshAllCommand(scopeId, locale);
+            return;
+        }
         if (action === 'use') {
             await this.handleAuthUseCommand(scopeId, locale, args.slice(1));
             return;
@@ -3889,6 +3910,26 @@ export class BridgeSessionCore {
         const messageId = await this.sendMessage(scopeId, renderAuthListMessage(locale, state, this.authDisplayBotLabel(), parseWeixinBridgeScope(scopeId) !== null), authChoiceKeyboard(locale, record));
         record.messageId = messageId;
     }
+    async handleAuthRefreshAllCommand(scopeId, locale) {
+        if (!this.canRunGlobalAuthRefresh()) {
+            await this.sendMessage(scopeId, t(locale, 'auth_refresh_all_blocked_active'));
+            return;
+        }
+        await this.sendMessage(scopeId, t(locale, 'auth_refresh_all_starting'));
+        const result = await this.refreshAllCodexAuthCandidates();
+        const state = await this.listCodexAuthState();
+        await this.applySharedCodexAuthQuotaSnapshots(state);
+        const record = {
+            localId: crypto.randomBytes(8).toString('hex'),
+            chatId: scopeId,
+            messageId: null,
+            candidates: state.candidates,
+            createdAt: Date.now(),
+        };
+        this.pendingAuthChoiceLists.set(record.localId, record);
+        const messageId = await this.sendMessage(scopeId, `${formatAuthRefreshAllResult(locale, result)}\n\n${renderAuthListMessage(locale, state, this.authDisplayBotLabel(), parseWeixinBridgeScope(scopeId) !== null)}`, authChoiceKeyboard(locale, record));
+        record.messageId = messageId;
+    }
     async handleAuthUseCommand(scopeId, locale, args) {
         if (this.hasLocalBlockingActivity()) {
             await this.sendMessage(scopeId, t(locale, 'auth_reload_blocked_active'));
@@ -4361,6 +4402,25 @@ export class BridgeSessionCore {
             await this.handleLoginDeviceCommand(event.scopeId, locale);
             return;
         }
+        if (action === 'refresh_all') {
+            if (!this.canRunGlobalAuthRefresh()) {
+                await this.messaging.answerCallback(event.callbackQueryId, t(locale, 'auth_refresh_all_blocked_active'));
+                return;
+            }
+            await this.messaging.answerCallback(event.callbackQueryId, t(locale, 'auth_refresh_all_starting'));
+            if (record.messageId !== null) {
+                await this.editMessage(event.scopeId, record.messageId, t(locale, 'auth_refresh_all_starting'), []);
+            }
+            const result = await this.refreshAllCodexAuthCandidates();
+            const state = await this.listCodexAuthState();
+            await this.applySharedCodexAuthQuotaSnapshots(state);
+            record.candidates = state.candidates;
+            record.createdAt = Date.now();
+            if (record.messageId !== null) {
+                await this.editMessage(event.scopeId, record.messageId, `${formatAuthRefreshAllResult(locale, result)}\n\n${renderAuthListMessage(locale, state, this.authDisplayBotLabel(), parseWeixinBridgeScope(event.scopeId) !== null)}`, authChoiceKeyboard(locale, record));
+            }
+            return;
+        }
         await this.messaging.answerCallback(event.callbackQueryId, t(locale, 'auth_reload_restarting'));
         await this.handleAuthReloadCommand(event.scopeId, locale);
     }
@@ -4603,6 +4663,71 @@ export class BridgeSessionCore {
             });
         }
     }
+    async refreshAllCodexAuthCandidates() {
+        if (this.authRefreshAllInProgress) {
+            throw new UserFacingError('Auth refresh all is already running.');
+        }
+        const initialState = await this.listCodexAuthState();
+        const result = { refreshed: [], skipped: [], failed: [] };
+        if (initialState.candidates.length === 0) {
+            return result;
+        }
+        const authStat = await fs.lstat(initialState.authPath).catch(() => null);
+        const originalRegularAuth = authStat?.isFile()
+            ? await fs.readFile(initialState.authPath, 'utf8').catch(() => null)
+            : null;
+        let changedAuthTarget = false;
+        this.authRefreshAllInProgress = true;
+        try {
+            for (const candidate of initialState.candidates) {
+                const before = await readChatGptAuthMetadata(candidate.path);
+                if (!before) {
+                    result.skipped.push(candidate.name);
+                    continue;
+                }
+                try {
+                    await pointCodexAuthAtTarget(initialState.authDir, initialState.authPath, candidate.path);
+                    changedAuthTarget = true;
+                    this.pendingTurnErrors.clear();
+                    this.attachedThreads.clear();
+                    await this.app.restart();
+                    await this.app.readAccount(true);
+                    const rateLimits = await this.app.readAccountRateLimits();
+                    const snapshot = selectCodexRateLimitSnapshot(rateLimits);
+                    if (!snapshot) {
+                        throw new Error('Codex did not return ChatGPT rate limits after refresh');
+                    }
+                    const after = await readChatGptAuthMetadata(candidate.path);
+                    if (!after || after.accountId !== before.accountId) {
+                        throw new Error('refreshed auth account id did not match the original candidate');
+                    }
+                    if (after.lastRefreshMs <= before.lastRefreshMs) {
+                        throw new Error('Codex did not advance last_refresh');
+                    }
+                    await this.recordCodexAuthQuotaSnapshot(candidate.name, after.accountId, snapshot);
+                    await this.syncCodexAuthCandidate(candidate.name);
+                    result.refreshed.push(candidate.name);
+                }
+                catch (error) {
+                    result.failed.push({ name: candidate.name, error: formatUserError(error) });
+                }
+            }
+        }
+        finally {
+            try {
+                if (changedAuthTarget) {
+                    await restoreCodexAuthTarget(initialState.authDir, initialState.authPath, initialState.currentTargetPath, originalRegularAuth);
+                    this.pendingTurnErrors.clear();
+                    this.attachedThreads.clear();
+                    await this.app.restart();
+                }
+            }
+            finally {
+                this.authRefreshAllInProgress = false;
+            }
+        }
+        return result;
+    }
     async buildNativeCollaborationMode(settings, cwd, modeOverride) {
         const mode = resolveCollaborationMode(modeOverride === undefined ? settings?.collaborationMode ?? null : modeOverride);
         try {
@@ -4807,11 +4932,7 @@ export class BridgeSessionCore {
             }
             const quota = authQuotaSnapshotFromRateLimit(snapshot, metadata?.accountId ?? null);
             candidate.quota = quota;
-            this.authQuotaSnapshots[candidate.name] = quota;
-            if (metadata?.accountId) {
-                this.store.setCodexAuthQuotaSnapshot(this.authRuntimeId(), candidate.name, metadata.accountId, quota);
-            }
-            await this.writeCodexAuthQuotaSnapshots();
+            await this.recordCodexAuthQuotaSnapshot(candidate.name, metadata?.accountId ?? null, snapshot);
             await this.applySharedCodexAuthQuotaSnapshots(state);
             await this.syncCodexAuthCandidate(candidate.name);
         }
@@ -4840,6 +4961,14 @@ export class BridgeSessionCore {
             candidate.quota = mergeCodexAuthQuotaSnapshots(candidate.quota, snapshotsByAccount.get(accountId) ?? null);
         }
     }
+    async recordCodexAuthQuotaSnapshot(candidateName, accountId, snapshot) {
+        const quota = authQuotaSnapshotFromRateLimit(snapshot, accountId);
+        this.authQuotaSnapshots[candidateName] = quota;
+        if (accountId) {
+            this.store.setCodexAuthQuotaSnapshot(this.authRuntimeId(), candidateName, accountId, quota);
+        }
+        await this.writeCodexAuthQuotaSnapshots();
+    }
     async readCodexAuthCandidateAccountIds(candidates) {
         const entries = await Promise.all(candidates.map(async (candidate) => {
             const metadata = await readChatGptAuthMetadata(candidate.path);
@@ -7455,6 +7584,30 @@ async function pointCodexAuthAtTarget(authDir, authPath, targetPath) {
         throw error;
     }
 }
+async function restoreCodexAuthTarget(authDir, authPath, targetPath, regularAuthContents) {
+    if (regularAuthContents !== null) {
+        await fs.mkdir(authDir, { recursive: true });
+        const temporary = path.join(authDir, `.auth.json.${process.pid}.${Date.now()}.restore`);
+        try {
+            await fs.writeFile(temporary, regularAuthContents, { mode: 0o600 });
+            await fs.rename(temporary, authPath);
+        }
+        catch (error) {
+            await fs.unlink(temporary).catch(() => undefined);
+            throw error;
+        }
+        return;
+    }
+    if (targetPath) {
+        await pointCodexAuthAtTarget(authDir, authPath, targetPath);
+        return;
+    }
+    await fs.unlink(authPath).catch((error) => {
+        if (!isFileMissingError(error)) {
+            throw error;
+        }
+    });
+}
 async function switchCodexAuth(targetPath, explicitAuthDir = null) {
     const state = await listCodexAuthState(new Set(), explicitAuthDir);
     const candidate = state.candidates.find(entry => entry.path === targetPath);
@@ -7513,9 +7666,33 @@ function authChoiceKeyboard(locale, record) {
         { text: t(locale, 'button_permissions'), callback_data: 'nav:permissions' },
         { text: t(locale, 'button_login_device'), callback_data: `auth:${record.localId}:login_device` },
     ]);
-    rows.push([{ text: t(locale, 'button_auth_reload'), callback_data: `auth:${record.localId}:reload` }]);
+    rows.push([
+        { text: t(locale, 'button_auth_reload'), callback_data: `auth:${record.localId}:reload` },
+        { text: t(locale, 'button_auth_refresh_all'), callback_data: `auth:${record.localId}:refresh_all` },
+    ]);
     return rows;
 }
+function formatAuthRefreshAllResult(locale, result) {
+    const lines = [t(locale, 'auth_refresh_all_done', {
+            refreshed: String(result.refreshed.length),
+            skipped: String(result.skipped.length),
+            failed: String(result.failed.length),
+        })];
+    if (result.refreshed.length > 0) {
+        lines.push(t(locale, 'auth_refresh_all_refreshed', { value: result.refreshed.join(', ') }));
+    }
+    if (result.skipped.length > 0) {
+        lines.push(t(locale, 'auth_refresh_all_skipped', { value: result.skipped.join(', ') }));
+    }
+    if (result.failed.length > 0) {
+        const details = result.failed
+            .slice(0, 5)
+            .map((failure) => `${failure.name}: ${failure.error}`)
+            .join('; ');
+        lines.push(t(locale, 'auth_refresh_all_failed', { value: details }));
+    }
+    return lines.join('\n');
+}
 function normalizeHelpUsageKey(name) {
     const normalized = name.toLowerCase();
     switch (normalized) {

package/dist/i18n.d.ts

@@ -137,7 +137,7 @@ declare const MESSAGES: {
         readonly auth_reload_restarting: "Restarting Codex app-server to reload auth...";
         readonly auth_reload_done: "Codex app-server restarted. Current auth has been reloaded.";
         readonly auth_reload_blocked_active: "Cannot reload Codex auth while a turn, approval, or question is active. Wait or use /interrupt first.";
-        readonly usage_auth: "Usage: /auth [list|use <n>|enable <n>|disable <n>|reload|add <name>]";
+        readonly usage_auth: "Usage: /auth [list|use <n>|enable <n>|disable <n>|reload|refresh all|add <name>]";
         readonly usage_auth_add: "Usage: /auth add <name>. Use letters, numbers, dot, dash, or underscore.";
         readonly auth_list_title: "Codex auth files:";
         readonly auth_bot: "Bot runtime: {value}";
@@ -158,6 +158,12 @@ declare const MESSAGES: {
         readonly auth_switching: "Switching Codex auth: {from} -> {to}...";
         readonly auth_switch_done: "Codex auth switched: {from} -> {to}.";
         readonly auth_recovered_newer_candidate: "Recovered a newer same-account credential for {value} from another Codex home before reload.";
+        readonly auth_refresh_all_starting: "Refreshing all ChatGPT auth candidates. Every runtime must stay idle...";
+        readonly auth_refresh_all_blocked_active: "Cannot refresh all auth candidates while any runtime, approval, input, login, or auth mirror write is active. Wait or use /interrupt first.";
+        readonly auth_refresh_all_done: "Auth refresh all complete: {refreshed} refreshed, {skipped} skipped, {failed} failed.";
+        readonly auth_refresh_all_refreshed: "Refreshed: {value}";
+        readonly auth_refresh_all_skipped: "Skipped non-ChatGPT/invalid candidates: {value}";
+        readonly auth_refresh_all_failed: "Failed: {value}";
         readonly auth_auto_switching: "Codex auth problem detected ({error}). Switching: {from} -> {to}...";
         readonly auth_auto_done: "Auto-switched Codex auth: {from} -> {to}.";
         readonly auth_auto_retrying: "Retrying the same request with the new auth...";
@@ -173,6 +179,7 @@ declare const MESSAGES: {
         readonly auth_add_missing_file: "Login completed, but the new auth file was not created: {value}";
         readonly button_login_device: "🔑 Login";
         readonly button_auth_reload: "🔄 Reload auth";
+        readonly button_auth_refresh_all: "🔁 Refresh all";
         readonly button_auth_enable: "✅";
         readonly button_auth_disable: "⏸️";
         readonly another_turn_running: "Another turn is already running. Use /interrupt, /takeover, /queue, or wait.";
@@ -713,7 +720,7 @@ declare const MESSAGES: {
         readonly auth_reload_restarting: "正在重启 Codex app-server 以重新读取 auth...";
         readonly auth_reload_done: "Codex app-server 已重启，当前 auth 已重新读取。";
         readonly auth_reload_blocked_active: "当前有回复、审批或问题在进行中，不能重载 Codex auth。请先等待，或使用 /interrupt。";
-        readonly usage_auth: "用法：/auth [list|use <编号>|enable <编号>|disable <编号>|reload|add <名称>]";
+        readonly usage_auth: "用法：/auth [list|use <编号>|enable <编号>|disable <编号>|reload|refresh all|add <名称>]";
         readonly usage_auth_add: "用法：/auth add <名称>。名称只能包含字母、数字、点、短横线或下划线。";
         readonly auth_list_title: "Codex auth 文件：";
         readonly auth_bot: "Bot runtime：{value}";
@@ -734,6 +741,12 @@ declare const MESSAGES: {
         readonly auth_switching: "正在切换 Codex auth：{from} -> {to}...";
         readonly auth_switch_done: "Codex auth 已切换：{from} -> {to}。";
         readonly auth_recovered_newer_candidate: "重载前已从其他 Codex home 恢复 {value} 的同账号较新凭据。";
+        readonly auth_refresh_all_starting: "正在刷新全部 ChatGPT auth 候选。执行期间所有 runtime 必须保持空闲...";
+        readonly auth_refresh_all_blocked_active: "当前有任一 runtime、审批、待输入、登录或 auth 镜像写入在进行中，不能刷新全部 auth。请先等待，或使用 /interrupt。";
+        readonly auth_refresh_all_done: "全部 auth 刷新完成：已刷新 {refreshed}，已跳过 {skipped}，失败 {failed}。";
+        readonly auth_refresh_all_refreshed: "已刷新：{value}";
+        readonly auth_refresh_all_skipped: "已跳过非 ChatGPT/无效候选：{value}";
+        readonly auth_refresh_all_failed: "失败：{value}";
         readonly auth_auto_switching: "检测到 Codex auth 问题（{error}），正在切换：{from} -> {to}...";
         readonly auth_auto_done: "已自动切换 Codex auth：{from} -> {to}。";
         readonly auth_auto_retrying: "正在用新的 auth 重试同一条请求...";
@@ -749,6 +762,7 @@ declare const MESSAGES: {
         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_refresh_all: "🔁 刷新全部";
         readonly button_auth_enable: "✅";
         readonly button_auth_disable: "⏸️";
         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。";

package/dist/i18n.js

@@ -135,7 +135,7 @@ const MESSAGES = {
         auth_reload_restarting: 'Restarting Codex app-server to reload auth...',
         auth_reload_done: 'Codex app-server restarted. Current auth has been reloaded.',
         auth_reload_blocked_active: 'Cannot reload Codex auth while a turn, approval, or question is active. Wait or use /interrupt first.',
-        usage_auth: 'Usage: /auth [list|use <n>|enable <n>|disable <n>|reload|add <name>]',
+        usage_auth: 'Usage: /auth [list|use <n>|enable <n>|disable <n>|reload|refresh all|add <name>]',
         usage_auth_add: 'Usage: /auth add <name>. Use letters, numbers, dot, dash, or underscore.',
         auth_list_title: 'Codex auth files:',
         auth_bot: 'Bot runtime: {value}',
@@ -156,6 +156,12 @@ const MESSAGES = {
         auth_switching: 'Switching Codex auth: {from} -> {to}...',
         auth_switch_done: 'Codex auth switched: {from} -> {to}.',
         auth_recovered_newer_candidate: 'Recovered a newer same-account credential for {value} from another Codex home before reload.',
+        auth_refresh_all_starting: 'Refreshing all ChatGPT auth candidates. Every runtime must stay idle...',
+        auth_refresh_all_blocked_active: 'Cannot refresh all auth candidates while any runtime, approval, input, login, or auth mirror write is active. Wait or use /interrupt first.',
+        auth_refresh_all_done: 'Auth refresh all complete: {refreshed} refreshed, {skipped} skipped, {failed} failed.',
+        auth_refresh_all_refreshed: 'Refreshed: {value}',
+        auth_refresh_all_skipped: 'Skipped non-ChatGPT/invalid candidates: {value}',
+        auth_refresh_all_failed: 'Failed: {value}',
         auth_auto_switching: 'Codex auth problem detected ({error}). Switching: {from} -> {to}...',
         auth_auto_done: 'Auto-switched Codex auth: {from} -> {to}.',
         auth_auto_retrying: 'Retrying the same request with the new auth...',
@@ -171,6 +177,7 @@ const MESSAGES = {
         auth_add_missing_file: 'Login completed, but the new auth file was not created: {value}',
         button_login_device: '🔑 Login',
         button_auth_reload: '🔄 Reload auth',
+        button_auth_refresh_all: '🔁 Refresh all',
         button_auth_enable: '✅',
         button_auth_disable: '⏸️',
         another_turn_running: 'Another turn is already running. Use /interrupt, /takeover, /queue, or wait.',
@@ -711,7 +718,7 @@ const MESSAGES = {
         auth_reload_restarting: '正在重启 Codex app-server 以重新读取 auth...',
         auth_reload_done: 'Codex app-server 已重启，当前 auth 已重新读取。',
         auth_reload_blocked_active: '当前有回复、审批或问题在进行中，不能重载 Codex auth。请先等待，或使用 /interrupt。',
-        usage_auth: '用法：/auth [list|use <编号>|enable <编号>|disable <编号>|reload|add <名称>]',
+        usage_auth: '用法：/auth [list|use <编号>|enable <编号>|disable <编号>|reload|refresh all|add <名称>]',
         usage_auth_add: '用法：/auth add <名称>。名称只能包含字母、数字、点、短横线或下划线。',
         auth_list_title: 'Codex auth 文件：',
         auth_bot: 'Bot runtime：{value}',
@@ -732,6 +739,12 @@ const MESSAGES = {
         auth_switching: '正在切换 Codex auth：{from} -> {to}...',
         auth_switch_done: 'Codex auth 已切换：{from} -> {to}。',
         auth_recovered_newer_candidate: '重载前已从其他 Codex home 恢复 {value} 的同账号较新凭据。',
+        auth_refresh_all_starting: '正在刷新全部 ChatGPT auth 候选。执行期间所有 runtime 必须保持空闲...',
+        auth_refresh_all_blocked_active: '当前有任一 runtime、审批、待输入、登录或 auth 镜像写入在进行中，不能刷新全部 auth。请先等待，或使用 /interrupt。',
+        auth_refresh_all_done: '全部 auth 刷新完成：已刷新 {refreshed}，已跳过 {skipped}，失败 {failed}。',
+        auth_refresh_all_refreshed: '已刷新：{value}',
+        auth_refresh_all_skipped: '已跳过非 ChatGPT/无效候选：{value}',
+        auth_refresh_all_failed: '失败：{value}',
         auth_auto_switching: '检测到 Codex auth 问题（{error}），正在切换：{from} -> {to}...',
         auth_auto_done: '已自动切换 Codex auth：{from} -> {to}。',
         auth_auto_retrying: '正在用新的 auth 重试同一条请求...',
@@ -747,6 +760,7 @@ const MESSAGES = {
         auth_add_missing_file: '登录已完成，但没有创建新的 auth 文件：{value}',
         button_login_device: '🔑 设备登录',
         button_auth_reload: '🔄 重载 auth',
+        button_auth_refresh_all: '🔁 刷新全部',
         button_auth_enable: '✅',
         button_auth_disable: '⏸️',
         another_turn_running: '已经有一个回复在进行中。请先等待，或使用 /interrupt、/takeover、/queue。',

package/docs/user-manual.md

@@ -389,7 +389,7 @@ If the login is cancelled or fails, FoxClaw tries to restore the previous auth t
 
 ### 6.3 The `/auth` Panel
 
-`/auth` lists candidate accounts, the current account, and the auth directory. It also provides buttons for switching, disabling, login, and reload. In multi-bot mode the panel names the `@botname` runtime being managed, because private chats, groups, and topics on one bot share that bot's current auth. The `5h|7d` numbers before each filename are the last recorded remaining percentages for the two quota windows; the current auth is refreshed when the panel opens, while other candidates are not switched merely to query quota. When multiple bot runtimes have recently used the same ChatGPT account, FoxClaw combines their cached quota snapshots by verified account ID, so one bot's `/auth` panel can show quota information learned by another bot without mixing different accounts.
+`/auth` lists candidate accounts, the current account, and the auth directory. It also provides buttons for switching, disabling, login, reload, and Refresh all. In multi-bot mode the panel names the `@botname` runtime being managed, because private chats, groups, and topics on one bot share that bot's current auth. The `5h|7d` numbers before each filename are the last recorded remaining percentages for the two quota windows; the current auth is refreshed when the panel opens, while other candidates are not switched merely to query quota. When multiple bot runtimes have recently used the same ChatGPT account, FoxClaw combines their cached quota snapshots by verified account ID, so one bot's `/auth` panel can show quota information learned by another bot without mixing different accounts.
 
 Approximation:
 
@@ -405,10 +405,10 @@ Quota remaining: 5h|7d|auth
 [✅ 20|25|auth.json_personal] [✅]
 [🔐 --|--|auth.json_team]     [✅]
 [🛡️ Access]             [🔑 Login]
-[🔄 Reload auth]
+[🔄 Reload auth]        [🔁 Refresh all]
 ```
 
-The right-side `✅` / `⏸️` button shows the current state. Tapping it toggles enabled/disabled, and the refreshed list shows the new state. Tapping a candidate switches auth, restarts that runtime, and refreshes the same panel with its buttons intact so you can switch again immediately. `--|--` means no quota snapshot has been observed for that candidate yet.
+The right-side `✅` / `⏸️` button shows the current state. Tapping it toggles enabled/disabled, and the refreshed list shows the new state. Tapping a candidate switches auth, restarts that runtime, and refreshes the same panel with its buttons intact so you can switch again immediately. `Refresh all` is a maintenance action: it is allowed only when every Telegram runtime, the Weixin runtime, approvals, inputs, logins, and auth mirroring are idle. It then visits every ChatGPT candidate in the panel, asks Codex to force-refresh tokens with `account/read refreshToken=true`, verifies the result through the usage endpoint, mirrors successful candidates, restores the original current auth, and refreshes the panel with a summary. `--|--` means no quota snapshot has been observed for that candidate yet.
 
 Equivalent commands:
 
@@ -417,6 +417,7 @@ Equivalent commands:
 - `/auth enable <n>`: let candidate n participate in auto-rotation.
 - `/auth disable <n>`: skip candidate n during auto-rotation.
 - `/auth reload` or `/auth_reload`: restart app-server and reload the current `auth.json`.
+- `/auth refresh all`: same as the panel's Refresh all button.
 
 If the requesting bot runtime has active turns, pending approvals, pending user inputs, or MCP elicitations, FoxClaw refuses manual auth switching to avoid changing accounts mid-request; another idle bot is unaffected.
 

package/docs/zh/user-manual.md

@@ -389,7 +389,7 @@ cp -L ~/.codex/auth.json ~/.codex/auth.json_personal
 
 ### 6.3 `/auth` 面板
 
-`/auth` 会列出候选账号、当前账号和 auth 目录，并提供按钮切换、禁用、登录和重载。多 bot 模式中，面板顶部还会显示当前正在管理的 `@botname`，因为该 bot 内的私聊、群聊和话题共享同一个当前 auth。每个候选名前的 `5h|7d` 数字表示上次记录到的两个额度窗口剩余百分比；当前 auth 会在打开面板时刷新，其他候选不会为了查询额度被自动切换。如果多个 bot runtime 最近使用过同一个 ChatGPT 账号，FoxClaw 会按已验证的账号 ID 合并它们缓存到的额度快照，因此一个 bot 的 `/auth` 面板可以显示另一个 bot 掌握到的额度信息，同时不会把不同账号混在一起。
+`/auth` 会列出候选账号、当前账号和 auth 目录，并提供按钮切换、禁用、登录、重载和刷新全部。多 bot 模式中，面板顶部还会显示当前正在管理的 `@botname`，因为该 bot 内的私聊、群聊和话题共享同一个当前 auth。每个候选名前的 `5h|7d` 数字表示上次记录到的两个额度窗口剩余百分比；当前 auth 会在打开面板时刷新，其他候选不会为了查询额度被自动切换。如果多个 bot runtime 最近使用过同一个 ChatGPT 账号，FoxClaw 会按已验证的账号 ID 合并它们缓存到的额度快照，因此一个 bot 的 `/auth` 面板可以显示另一个 bot 掌握到的额度信息，同时不会把不同账号混在一起。
 
 示意：
 
@@ -405,10 +405,10 @@ Candidates: 2
 [✅ 20|25|auth.json_personal] [✅]
 [🔐 --|--|auth.json_team]     [✅]
 [🛡️ Access]             [🔑 设备登录]
-[🔄 Reload auth]
+[🔄 Reload auth]        [🔁 刷新全部]
 ```
 
-右侧 `✅` / `⏸️` 表示当前状态。点一下会切换启用/禁用，列表刷新后图标会随状态变化。点击候选会切换 auth、重启对应 runtime，并在原消息上刷新面板且保留按钮，因此可以立即连续切换。`--|--` 表示该候选还没有额度历史快照。
+右侧 `✅` / `⏸️` 表示当前状态。点一下会切换启用/禁用，列表刷新后图标会随状态变化。点击候选会切换 auth、重启对应 runtime，并在原消息上刷新面板且保留按钮，因此可以立即连续切换。`刷新全部` 是维护操作：只有所有 Telegram runtime、微信 runtime、审批、待输入、登录流程和 auth 镜像写入都空闲时才允许执行。它会逐个访问面板里的 ChatGPT 候选，让 Codex 通过 `account/read refreshToken=true` 强制刷新 token，再用 usage 接口验证，成功后镜像到其他 bot home，最后恢复原本的当前 auth 并刷新面板摘要。`--|--` 表示该候选还没有额度历史快照。
 
 命令等价用法：
 
@@ -417,6 +417,7 @@ Candidates: 2
 - `/auth enable <n>`：让第 n 个候选参与自动轮转。
 - `/auth disable <n>`：禁用第 n 个候选，自动轮转会跳过它。
 - `/auth reload` 或 `/auth_reload`：重启 app-server，重新加载当前 `auth.json`。
+- `/auth refresh all`：等同于面板上的“刷新全部”按钮。
 
 切换 auth 时，如果当前 bot runtime 还有活跃 turn、待审批、待用户输入或 MCP elicitation，FoxClaw 会先拒绝切换，避免中途换号破坏正在进行的请求；另一个空闲 bot 不受影响。
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@foxden-app/foxclaw",
-  "version": "0.4.5",
+  "version": "0.4.7",
   "description": "Foxden local execution claw for controlling Codex from trusted chat interfaces.",
   "type": "module",
   "main": "dist/main.js",
@@ -15,6 +15,7 @@
     ".env.example",
     "README.md",
     "README_EN.md",
+    "CHANGELOG.md",
     "LICENSE"
   ],
   "private": false,

package/skills/foxclaw/SKILL.md

@@ -67,6 +67,7 @@ Telegram behavior:
 - when `TG_BOT_TOKENS` contains multiple bots, one FoxClaw service starts independent Codex app-servers and auth selections per bot
 - isolated Telegram runtimes use independent `CODEX_HOME` directories and file-backed credentials; they do not share Codex sessions
 - refreshed ChatGPT candidates are mirrored only after online usage-endpoint validation; before auth switch or reload, a runtime restores any newer same-account credential found in another Codex home
+- `/auth` includes a Refresh all panel button; it is a global-idle maintenance action that force-refreshes every ChatGPT candidate through Codex `account/read refreshToken=true`, validates usage, mirrors successful candidates, and restores the original current auth
 - if `TG_BOT_TOKEN` is also set to one exact token from `TG_BOT_TOKENS`, that bot uses the default/shared-terminal runtime instead of an isolated Telegram home
 - if Weixin is enabled alongside multiple Telegram bots, it remains on the default Codex runtime instead of borrowing a Telegram bot runtime
 - in a group with multiple configured bots, address a bot by mention, reply, or suffixed command such as `/status@botname`
@@ -173,6 +174,8 @@ Do not describe the setup as "done" until this smoke test has either passed or b
 
 Use this checklist when the user asks for standard closing actions, release wrap-up, local install updates, npm publish, or says things like "收尾动作", "中文 commit msg", "中英双语 commit msg", "push", "npm publish", or "本地安装更新".
 
+When the user says development is complete, asks to "收尾", or asks to push/publish, treat that as authorization to finish the whole release path without asking for another confirmation. After validation passes, commit the scoped changes, push the branch, publish through the configured release flow, and verify the published result. Stop and ask only when a real external decision is required, such as npm/GitHub web auth, missing credentials, failed checks, merge conflicts, an already-published target version that requires choosing a new version, or unrelated local changes that would be staged.
+
 1. Inspect scope before staging:
    - `git status -sb`
    - `git diff --stat`
@@ -188,6 +191,7 @@ Use this checklist when the user asks for standard closing actions, release wrap
    - For this repo, prefer a bilingual one-line subject in the form `中文 | English`.
    - Put Chinese first, keep the English half semantically equivalent, and keep both halves concise.
    - For release commits, use the same format, for example `发布 0.4.0：支持多机器人隔离 | Release 0.4.0: support multi-bot isolation`.
+   - Update `CHANGELOG.md` for any published version, using bilingual Chinese and English notes.
    - Never stage unrelated local changes.
 5. Push the current branch after a successful commit:
    - `git push origin <branch>`
@@ -202,12 +206,14 @@ Use this checklist when the user asks for standard closing actions, release wrap
    - `/update` now attempts to update globally npm/pnpm-managed Codex CLI installations and refuses to restart while any configured Telegram runtime, enabled Weixin default runtime, or auth mirror write is busy.
 7. Publish to npm when requested:
    - Prefer GitHub Actions trusted publishing via `.github/workflows/publish.yml`: bump and commit the package version, push `main`, then push a matching `v<version>` tag. The tag version must match `package.json`.
+   - Ensure the matching changelog section exists before pushing the version tag; the publish workflow uses it to create or update the GitHub Release.
    - Treat `workflow_dispatch` only as a retry path from an existing matching release tag; do not manually run publishing from `main`.
    - Configure npmjs.com trusted publishing for `foxden-app/foxclaw` and workflow `publish.yml`; do not store npm tokens if OIDC trusted publishing is available.
    - Temporary fallback: store an npm automation/bypass-2FA token as the GitHub Actions secret `NPM_TOKEN`. Never print the token or commit it.
    - Check `npm whoami` and `npm view @foxden-app/foxclaw version`.
    - If the target version is already published, bump with `npm version patch --no-git-tag-version` before validation and commit.
    - If a workflow fails before `npm publish`, inspect that failed GitHub Actions step before attributing it to npm trusted publishing.
+   - After publishing, verify both `npm view @foxden-app/foxclaw version` and the GitHub Release page for the pushed tag.
    - Manual fallback only: run `BROWSER=true npm publish` in a TTY.
    - If npm returns `ENEEDAUTH`, report that npm publish is blocked by registry login and leave the package unreported as published.
    - Never print npm tokens or `.npmrc` auth values.

package/skills/npm-publish/SKILL.md

@@ -7,6 +7,8 @@ description: Publish npm packages safely, especially packages that require npm 2
 
 Use this skill to publish an npm package from a repo. Prefer CI trusted publishing when the repository has a publish workflow. Prefer the normal no-2FA publish path first only when CI publishing is not available. Fall back to npm web-auth only when npm explicitly prompts for it.
 
+If the user asks to release, publish, "收尾", "push publish", or otherwise finish a completed package change, proceed through verification, commit, push, publish, and post-publish verification without asking for a second confirmation. Pause only for real blockers: missing credentials, failed checks, unrelated changes that would be staged, merge conflicts, an already-published target version that needs a version choice, or npm/GitHub web-auth that requires the user to click/confirm an external page.
+
 ## Release Checklist
 
 1. Sync first when the user asks for it:

package/skills/npm-publish/agents/openai.yaml

@@ -1,4 +1,4 @@
 interface:
   display_name: "NPM Publish"
-  short_description: "Publish npm packages with 2FA web auth"
-  default_prompt: "Publish this npm package using the recorded web-auth flow."
+  short_description: "Verify, push, and publish npm packages"
+  default_prompt: "Publish this npm package: verify it, commit and push scoped release changes, run the configured publish flow, and only pause for real auth or failure blockers."