lark-remote 0.1.13 → 0.2.1

package/CHANGELOG.md

@@ -5,6 +5,41 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.2.1] - 2026-09-11
+
+### 新增
+
+- **Windows 正式支持**：原生 Windows 支持已在真机完成验证，README / getting-started 的平台前置条件从「macOS / Linux，暂不支持 Windows」更新为 macOS / Linux / Windows 10+
+
+### 变更
+
+- **安全文档措辞更新**：README 安全警告与 `SECURITY.md` 反映 Claude 审批模式——`permissionMode` 可配置（默认 `bypassPermissions`，配置为其他值即启用飞书卡片交互式审批），不再表述为「固定 bypassPermissions」；审批卡片描述从 Codex 专属改为 Claude / Codex / Kimi 通用
+
+## [0.2.0] - 2026-09-11
+
+### 新增
+
+- **Windows 平台支持（预发布）**：新增跨平台平台层（`src/platform/`）——路径、可执行文件发现、进程终止、spawn、shell 的抽象与 posix / win32 双实现；命令拉起统一走 `cross-spawn`，posix 行为零变化。Windows 尚未正式发布，README 前置条件已相应标注
+- **`/download <path>`（别名 `/d`）**：把本地文件直接发到会话，不经 `/ls` 卡片。路径规则同 `/ls`（`~` 展开、相对 cwd、30MB 上限），目录 / 不存在 / 超限给出明确错误
+- **`/ls <file>` 单文件卡**：传入文件路径渲染单文件卡（完整路径、大小、修改时间、「📎 下载」+「上级」按钮），不再报 `ls: xxx: Not a directory`
+- **分页栏页码跳转**：所有分页卡片（`/ls`、`/ws`、`/resume`、`/active`、`/order`）支持在分页栏输入页码直达；越界钳到末页，非数字输入报错且不刷新卡片
+- **扫码向导跨平台**：二维码渲染改用 `qrcode-generator`，Windows 全块渲染 + 静区 + 图片兜底
+
+### 修复
+
+- `/ls` 的「返回」改为回到浏览起点（`/ls <dir>` 指定的目录），不再跳回 workspace cwd；浏览起点随卡片 payload 传递，翻页 / 刷新 / 上级 / 目录按钮不会丢失
+- codex 会话兼容 rollout ≥0.153.4 的 `item_completed` 词表，修复 `/resume` 标题丢失
+- compact 卡片空内容渲染「Compact 完成」，不再显示通用空占位
+- 信息保真：notices 改为时间线消息渲染、省略提示固定置顶、工具卡片消除孤立代理项
+- Windows：隐藏 cmd.exe 控制台窗口（`windowsHide`）
+
+### 变更
+
+- **移除 `output.showThinking` / `showToolUse` / `showToolResult` 配置项与 `/config` 的「📤 输出」页签**：thinking、工具调用与工具结果恒常展示；旧配置文件的 `output` 段在加载时自动忽略，并在下次保存时从磁盘清理
+- 代码精简：ACP / translator / session 共享层下沉（`BaseAcpTranslator` / `BaseAcpRunner` / two-pass 扫描），死代码清理与平行重复收敛
+- 依赖：`qrcode-terminal` 换为 `qrcode-generator`；新增 `cross-spawn`（Windows 命令拉起）
+- 测试基建：新增 scoped typecheck（`tsconfig.test-infra.json`）覆盖 `tests/lib/**` 共享工厂；跨平台测试稳定化（win32 路径归一化、并行 vitest 内存上限）
+
 ## [0.1.13] - 2026-09-04
 
 ### 修复
@@ -37,7 +72,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### 变更
 
 - kimi ACP 能力声明：`fs` 保持关闭（走本地磁盘兜底，避免未实现的 `fs/*` RPC 报 `METHOD_NOT_FOUND`），`terminal` 打开
-- 文档清理：移除设计文档中对内部文档名（`lessons-redlines`）与本地仓库路径（`/Users/.../lark-remote`）的引用
 
 ## [0.1.10] - 2026-08-22
 
@@ -187,7 +221,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - `displayTitle`/`summary` 跳过 task-notification 注入
 - 修复 `scanClaudeSessions` 对迁移到其他目录的会话的遗漏（S2）
 - 修复 probe 生命周期、工厂回退、卡片回调等评审发现的问题（F1-F8）
-- 清理引用已删除文档（lessons-redlines、resume-card-budget-failure-analysis）导致的失效链接
 
 ### 变更
 
@@ -196,7 +229,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - 移除 `pnpm-lock.yaml`，统一使用 `bun.lock` 管理依赖
 - 新增 husky lint-staged pre-commit / pre-push 钩子（提交前 lint + typecheck）
 - 启用 npm provenance（构建产物可溯源）
-- 移除 watchdog 相关用户文档及 `open-source-playbook` 等临时文档
 
 ### 测试
 
@@ -275,3 +307,4 @@ Initial release.
 [0.1.6]: https://github.com/bungabungawoda/lark-remote/releases/tag/v0.1.6
 [0.1.7]: https://github.com/bungabungawoda/lark-remote/releases/tag/v0.1.7
 [0.1.8]: https://github.com/bungabungawoda/lark-remote/releases/tag/v0.1.8
+[0.2.1]: https://github.com/bungabungawoda/lark-remote/releases/tag/v0.2.1

package/README.en.md

@@ -16,20 +16,20 @@ Designed for a single-user, peer-to-peer private chat. The agent's system prompt
 
 ## ⚠️ Security Warning (read first)
 
-This tool turns Feishu messages directly into local agent executions. Claude runs with **full permissions** by default (fixed `bypassPermissions`); Codex uses the **app-server approval mode** (default `approvalPolicy=on-request` — command execution requires your confirmation on the Feishu card, sandbox defaults to `workspace-write`), which is relatively safe; other agents run with their CLI defaults. Agents can still read and write the local directories you specify. You MUST:
+This tool turns Feishu messages directly into local agent executions. **Out of the box, agents run without approval prompts**: Claude defaults to `bypassPermissions` (no approval card), but that permission mode is **configurable** — set it to anything other than `bypassPermissions` (`default` / `acceptEdits` / `auto` / `dontAsk` / `plan`) to enable interactive approval, where high-risk actions wait for your confirmation on the Feishu card. Codex uses the **app-server approval mode** (default `approvalPolicy=on-request` — command execution requires your confirmation on the Feishu card, sandbox defaults to `workspace-write`); Kimi defaults to `manual` (per-action approval); other agents run with their CLI defaults. **In any mode**, agents can read and write the local directories you specify. You MUST:
 
 - **Use it only in your own private (p2p) chat.** Never add the bot to any group chat; never let anyone else talk to it.
 - Restrict the app's visibility to **yourself only** in the Feishu Open Platform.
 - Don't run it on a machine holding data you can't afford to lose; prefer a dedicated user / machine / container.
 - Keep credentials such as `appSecret` in your local `config.yaml` only — **never commit them anywhere**.
 
-In Codex approval mode, an approval card arrives before any command runs — approve or reject it right in Feishu:
+In approval mode (Claude, Codex, Kimi), an approval card arrives before a command runs — approve or reject it right in Feishu:
 
-![Codex command approval card example](docs/images/approval-card.png)
+![Command approval card example](docs/images/approval-card.png)
 
 ## Prerequisites
 
-- OS: macOS / Linux. **Windows is not supported yet** (relies on POSIX behaviors such as Unix signals, bash, and file locks)
+- OS: macOS / Linux / Windows 10+ (native Windows support is verified on real hardware, with behavior matching macOS/Linux)
 - Node.js 20+ ([Bun](https://bun.sh/) for development)
 - A Feishu custom app (either way): scan-to-create (a QR code pops up in the terminal on first launch; scan it with the Feishu app to create the app and write credentials automatically); or create one manually in the Open Platform — enable bot capability, subscribe to `im.message.receive_v1` and `card.action.trigger` via **long connection** (WebSocket, no public address needed), with at least the `im:message` scope.
 - Claude Code CLI installed locally and logged in once in a terminal (`claude` → browser OAuth). Same idea for other agents (codex / opencode / pi / kimi): install the corresponding CLI first. DSH connects to a local DSH Web Host via HTTP+WebSocket (no local subprocess), default address `http://127.0.0.1:3080`.
@@ -70,14 +70,9 @@ defaultAgent: claude
 claude:
   model: claude-opus-4-8
   effort: medium            # low | medium | high | xhigh | max
-  # permissionMode is hardcoded to bypassPermissions (inside the runner)
+  permissionMode: bypassPermissions  # Claude official --permission-mode: default | acceptEdits | auto | bypassPermissions | manual | dontAsk | plan (switchable via /config card; anything other than bypassPermissions enables approval)
   stopGraceMs: 5000
 
-output:
-  showThinking: true
-  showToolUse: true
-  showToolResult: true
-
 logging:
   level: info               # debug | info | warn | error
 
@@ -113,7 +108,8 @@ Your original message also gets an emoji reaction: `Typing` while the run is in
 |---------|-------|----------|
 | `/help` | `/h` | Command list |
 | `/cd <path>` | - | Switch agent working directory (`~`, absolute/relative; clears the session) |
-| `/ls [dir]` | - | Directory/file card; click a directory to switch, click a file ≤30MB to send it to Feishu; paginates beyond 30 entries |
+| `/ls [dir\|file]` | - | Directory/file card; click a directory to browse, click a file ≤30MB to send it to Feishu; paginates beyond 30 entries with a page-number jump input. Passing a file path lists that file itself |
+| `/download <path>` | `/d` | Send a local file straight to the chat (`~`, absolute/relative paths, 30MB cap) |
 | `/ws save\|use\|remove` | - | Named directory aliases (`/ws` lists by default) |
 | `/resume [agent] [N\|id]` | `/r` | List/switch agent sessions for the current directory (card) |
 | `/active` | - | List all running sessions |

package/README.md

@@ -16,20 +16,20 @@
 
 ## ⚠️ 安全警告（必读）
 
-本工具会把飞书消息直接转为本地 agent 执行。Claude 默认以**完全权限**运行（固定 `bypassPermissions`）；Codex 采用 **app-server 审批模式**（默认 `approvalPolicy=on-request`，执行命令需在飞书卡片上确认，沙箱默认 `workspace-write`），相对安全；其他 agent 以各自 CLI 的默认权限运行。agent 仍可读写你指定的本地目录，请务必：
+本工具会把飞书消息直接转为本地 agent 执行。**开箱默认是不弹审批的**：Claude 默认以 `bypassPermissions` 运行（无审批卡），但该权限模式**可配置**——把它改成 `bypassPermissions` 以外的值（`default` / `acceptEdits` / `auto` / `dontAsk` / `plan`）即启用交互式审批，高风险操作会在飞书卡片上等你确认；Codex 采用 **app-server 审批模式**（默认 `approvalPolicy=on-request`，执行命令需在飞书卡片上确认，沙箱默认 `workspace-write`）；Kimi 默认 `manual`（逐条审批）；其他 agent 以各自 CLI 的默认权限运行。**无论哪种模式**，agent 都能读写你指定的本地目录，请务必：
 
 - **仅限你自己的私聊（p2p）使用**。不要把机器人拉进任何群聊，不要让任何其他人能与它对话。
 - 在飞书开放平台把应用可见范围限制到**只有你自己**。
 - 运行本工具的机器上不要存放与损失承受力不符的数据；建议在独立用户/机器/容器里运行。
 - `appSecret` 等凭据只在本地 `config.yaml`，**不要提交到任何仓库**。
 
-Codex 审批模式下，执行命令前会收到审批卡片，可直接在飞书里允许或拒绝：
+审批模式下（Claude、Codex、Kimi），执行命令前会收到审批卡片，可直接在飞书里允许或拒绝：
 
-![Codex 命令审批卡片示例](docs/images/approval-card.png)
+![命令审批卡片示例](docs/images/approval-card.png)
 
 ## 前置条件
 
-- 操作系统：macOS / Linux。**暂不支持 Windows**（依赖 Unix 信号、bash、文件锁等 POSIX 行为）
+- 操作系统：macOS / Linux / Windows 10+（原生 Windows 支持已完成真机验证，行为与 macOS/Linux 对齐）
 - Node.js 20+（开发时用 [Bun](https://bun.sh/)）
 - 飞书自建应用（二选一）：扫码创建（首次启动终端弹二维码，飞书 App 扫码即自动创建并写入凭据）；或手动在开放平台建应用——开启机器人能力，订阅 `im.message.receive_v1` 和 `card.action.trigger`，订阅方式选「长连接」（WebSocket，无需公网地址），权限至少 `im:message`。
 - Claude Code CLI：本地安装并在终端完成一次登录（`claude` → 浏览器 OAuth）。使用其他 agent（codex / opencode / pi / kimi）同理，先装好对应 CLI；DSH 需本地 DSH Web Host 在跑（默认 `http://127.0.0.1:3080`），lark-remote 通过 HTTP+WebSocket 直连、不 spawn 本地子进程。
@@ -70,14 +70,9 @@ defaultAgent: claude
 claude:
   model: claude-opus-4-8
   effort: medium            # low | medium | high | xhigh | max
-  # permissionMode 硬编码为 bypassPermissions（runner 内部）
+  permissionMode: bypassPermissions  # Claude 官方 --permission-mode：default | acceptEdits | auto | bypassPermissions | manual | dontAsk | plan（`/config` 卡片可切换；非 bypassPermissions 会启用审批）
   stopGraceMs: 5000
 
-output:
-  showThinking: true
-  showToolUse: true
-  showToolResult: true
-
 logging:
   level: info               # debug | info | warn | error
 
@@ -113,7 +108,8 @@ bridge 启动后不在终端输出，运行日志写入 `~/.lark-remote/logs/YYY
 |------|------|------|
 | `/help` | `/h` | 命令列表 |
 | `/cd <path>` | - | 切换 agent 工作目录（支持 `~`、绝对/相对路径，清空会话） |
-| `/ls [dir]` | - | 弹出目录/文件卡片；点击目录切换，点击 30MB 内文件发送到飞书；条目 >30 时支持翻页 |
+| `/ls [dir\|file]` | - | 弹出目录/文件卡片；点击目录切换，点击 30MB 内文件发送到飞书；条目 >30 时支持翻页与直接输入页码跳转。传文件路径时直接列出该文件（可下载） |
+| `/download <path>` | `/d` | 直接下载（发送）指定文件到飞书（支持 `~`、绝对/相对路径，上限 30MB） |
 | `/ws save\|use\|remove` | - | 命名目录别名管理（`/ws` 默认列出） |
 | `/resume [agent] [N\|id]` | `/r` | 列出/切换当前目录的 agent session（卡片） |
 | `/active` | - | 列出所有正在进行中的 session |

package/dist/bridge/final-usage.d.ts

@@ -0,0 +1,24 @@
+import type { SessionReaderRegistry } from '../session/index.js';
+import type { AgentKind } from '../runner/index.js';
+export declare function errorMessage(error: unknown): string;
+export type FinalUsageSnapshot = {
+    contextLength?: number;
+    contextLimit?: number;
+    compactCount?: number;
+    compactPreContextLength?: number;
+    cacheReadTokens?: number;
+    cacheCreationTokens?: number;
+    totalTokens?: number;
+    inputTokens?: number;
+    outputTokens?: number;
+    cumulativeTotalTokens?: number;
+    cumulativeInputTokens?: number;
+    cumulativeOutputTokens?: number;
+    cumulativeCacheReadTokens?: number;
+    cumulativeCacheCreationTokens?: number;
+};
+/**
+ * Read final usage (contextLength + compactCount + cache tokens) from the
+ * session jsonl. Called after a run finishes.
+ */
+export declare function resolveFinalUsage(registry: SessionReaderRegistry, defaultAgent: AgentKind, sessionId: string | undefined, cwd: string, agentKind?: AgentKind): FinalUsageSnapshot | undefined;

package/dist/bridge/final-usage.js

@@ -0,0 +1,55 @@
+/**
+ * Final usage resolution extracted from Bridge (2026-09 round-2
+ * simplification): after a run finishes, re-read the session transcript via
+ * the agent's session reader — live stream-json does not emit
+ * compact_boundary, so the live contextLength (fallback to result.usage
+ * input+output) is unreliable. The transcript jsonl is authoritative for
+ * postTokens + compact event count.
+ *
+ * Bridge state is passed in explicitly (registry + default agent); the
+ * function itself is stateless.
+ */
+import { getLogger } from '../logger/index.js';
+export function errorMessage(error) {
+    return error instanceof Error ? error.message : String(error);
+}
+/**
+ * Read final usage (contextLength + compactCount + cache tokens) from the
+ * session jsonl. Called after a run finishes.
+ */
+export function resolveFinalUsage(registry, defaultAgent, sessionId, cwd, agentKind = defaultAgent) {
+    if (!sessionId)
+        return undefined;
+    try {
+        const content = registry.get(agentKind).readSessionContent(sessionId, cwd);
+        if (!content.usage) {
+            // EnterWorktree relocate (2026-08-04): jsonl read silently returned no
+            // usage (e.g. transcript moved mid-session). Surface it so token-stat
+            // fallback to per-run live usage is visible in logs, not silent.
+            getLogger().warn(`[bridge] resolveFinalUsage: no usage from jsonl sessionId=${sessionId} cwd=${cwd} agent=${agentKind}, card falls back to per-run live usage`);
+        }
+        return content.usage
+            ? {
+                contextLength: content.usage.contextLength,
+                contextLimit: content.usage.contextLimit,
+                compactCount: content.usage.compactCount,
+                compactPreContextLength: content.usage.compactPreContextLength,
+                cacheReadTokens: content.usage.cacheReadTokens,
+                cacheCreationTokens: content.usage.cacheCreationTokens,
+                totalTokens: content.usage.totalTokens,
+                inputTokens: content.usage.inputTokens,
+                outputTokens: content.usage.outputTokens,
+                cumulativeTotalTokens: content.usage.cumulativeTotalTokens,
+                cumulativeInputTokens: content.usage.cumulativeInputTokens,
+                cumulativeOutputTokens: content.usage.cumulativeOutputTokens,
+                cumulativeCacheReadTokens: content.usage.cumulativeCacheReadTokens,
+                cumulativeCacheCreationTokens: content.usage.cumulativeCacheCreationTokens,
+            }
+            : undefined;
+    }
+    catch (err) {
+        getLogger().warn(`[bridge] resolveFinalUsage failed sessionId=${sessionId}: ${errorMessage(err)}`);
+        return undefined;
+    }
+}
+//# sourceMappingURL=final-usage.js.map

package/dist/bridge/index.d.ts

@@ -361,13 +361,6 @@ export declare class Bridge {
      * Handles fallback card sending, completion notification, and cleanup.
      */
     private finalizeRun;
-    /**
-     * Read final usage (contextLength + compactCount + cache tokens) from the session jsonl.
-     * Called after a run finishes: live stream-json does not emit compact_boundary,
-     * so the live contextLength (fallback to result.usage input+output) is
-     * unreliable. The jsonl is authoritative for postTokens + compact event count.
-     */
-    private resolveFinalUsage;
     /**
      * Send a notification card when a run completes.
      * This card contains the session summary and a button to resume the session.

package/dist/bridge/index.js

@@ -1,5 +1,6 @@
 import { randomUUID } from 'node:crypto';
 import { sleep } from '../common/sleep.js';
+import { resolveFinalUsage, errorMessage } from './final-usage.js';
 import { getLogger } from '../logger/index.js';
 import { RunCardSession } from '../card/run-card-session.js';
 import { renderRunCard } from '../card/run-renderer.js';
@@ -48,6 +49,13 @@ const TERMINAL_REACTION_EMOJI = {
 function terminalReactionEmoji(terminal) {
     return TERMINAL_REACTION_EMOJI[terminal] ?? 'Done';
 }
+/**
+ * W2.4 单点化：`'runCompact' in runner` 鸭子探测（compact 按钮门控，
+ * design doc §6.2-2）。codex app-server + kimi/opencode ACP runner 满足。
+ */
+function runnerHasRunCompact(runner) {
+    return ('runCompact' in runner && typeof runner.runCompact === 'function');
+}
 /**
  * Bridge owns the per-workspace serial processing queue (§9.6) and the
  * single-card streaming forwarding loop (§9.12 watchdog + session sync).
@@ -245,7 +253,7 @@ export class Bridge {
      */
     hasRunCompact(workspace, kind = this.config.defaultAgent) {
         const runner = this.getRunner(workspace, kind);
-        return ('runCompact' in runner && typeof runner.runCompact === 'function');
+        return runnerHasRunCompact(runner);
     }
     /** 入队时刻快照：当前 defaultAgent + 该 agent 的 sessionId（无 session 则 undefined）。
      *  唯一捕获点：排队消息在 T0 把 agent+session 钉进 AgentBinding，随任务闭包带到 T1
@@ -691,7 +699,7 @@ export class Bridge {
             const workspaces = this.workspaceStore.list();
             if (workspaces.length > 0) {
                 // NOTE: fallback uses insertion order, not sort preference — by design
-                // (workspace-sorting.md §9: cwd fallback stays insertion-order for now)
+                // (cwd fallback stays insertion-order for now)
                 cwd = workspaces[0].path;
                 getLogger().info(`[bridge] resolveCwd fallback cwd=${cwd} (first saved workspace)`);
             }
@@ -764,9 +772,6 @@ export class Bridge {
             replyTo: ctx.messageId,
             runId,
             renderOptions: {
-                showThinking: this.config.output.showThinking,
-                showToolUse: this.config.output.showToolUse,
-                showToolResult: this.config.output.showToolResult,
                 // Run card header shows the current agent name (e.g. "Claude · 思考中")
                 agentKind,
                 // Compact 按钮能力门控与 resume 卡 / auto-resume 卡一致（hasRunCompact
@@ -834,6 +839,8 @@ export class Bridge {
         let finalCumulativeOutputTokens;
         let finalCumulativeCacheReadTokens;
         let finalCumulativeCacheCreationTokens;
+        // Reasoning tokens (pi usage.reasoning) from the live result event.
+        let finalReasoningTokens;
         // 统一的 usage meta 构造（finalize 各分支共享，本方法内原先 5 处手写 11-15
         // 字段对象；第 6 处 streamCodexCompact 形状不同未收敛）。
         // catch 路径需覆盖 flow 字段时在展开后覆写即可。catch error 路径亦复用此
@@ -853,6 +860,7 @@ export class Bridge {
             cumulativeOutputTokens: finalCumulativeOutputTokens,
             cumulativeCacheReadTokens: finalCumulativeCacheReadTokens,
             cumulativeCacheCreationTokens: finalCumulativeCacheCreationTokens,
+            reasoningTokens: finalReasoningTokens,
         });
         try {
             try {
@@ -1047,6 +1055,9 @@ export class Bridge {
                         if (u.totalTokens !== undefined) {
                             finalTotalTokens = u.totalTokens;
                         }
+                        if (u.reasoningTokens !== undefined) {
+                            finalReasoningTokens = u.reasoningTokens;
+                        }
                     }
                 }
                 // Approval events (Codex app-server mode)
@@ -1125,7 +1136,7 @@ export class Bridge {
             const finalSessionId = this.sessionStore.getSessionId(ctx.userId, agentKind);
             const finalUsage = resultNotSuccess
                 ? undefined
-                : this.resolveFinalUsage(finalSessionId, cwd, agentKind);
+                : resolveFinalUsage(this.sessionReaderRegistry, this.config.defaultAgent, finalSessionId, cwd, agentKind);
             finalContextLength = finalUsage?.contextLength ?? contextLength;
             finalContextLimit = finalUsage?.contextLimit ?? liveContextLimit;
             finalCompactCount = finalUsage?.compactCount;
@@ -1243,9 +1254,6 @@ export class Bridge {
             if (settleResult === 'unsent') {
                 const sent = await this.sendResult({
                     card: renderRunCard(cardSession.currentState, {
-                        showThinking: this.config.output.showThinking,
-                        showToolUse: this.config.output.showToolUse,
-                        showToolResult: this.config.output.showToolResult,
                         agentKind: activeRun.agentKind,
                         // 与 createRunSession 的 renderOptions 一致：fallback 卡也要按
                         // runner 能力门控 Compact 按钮（缺省 undefined = 显示按钮，
@@ -1295,8 +1303,7 @@ export class Bridge {
                 // §6.2-3：写入条件从「codex 终态」放宽为「runner 有 runCompact 的终态」
                 // （codex app-server + kimi acp 均满足）。
                 const runTerminal = cardSession.currentState.terminal;
-                const runnerHasCompact = 'runCompact' in activeRun.runner &&
-                    typeof activeRun.runner.runCompact === 'function';
+                const runnerHasCompact = runnerHasRunCompact(activeRun.runner);
                 if (runnerHasCompact &&
                     activeRun.runner.lifetime === 'workspace' &&
                     (runTerminal === 'done' ||
@@ -1317,47 +1324,6 @@ export class Bridge {
             }
         }
     }
-    /**
-     * Read final usage (contextLength + compactCount + cache tokens) from the session jsonl.
-     * Called after a run finishes: live stream-json does not emit compact_boundary,
-     * so the live contextLength (fallback to result.usage input+output) is
-     * unreliable. The jsonl is authoritative for postTokens + compact event count.
-     */
-    resolveFinalUsage(sessionId, cwd, agentKind = this.config.defaultAgent) {
-        if (!sessionId)
-            return undefined;
-        try {
-            const content = this.sessionReaderRegistry.get(agentKind).readSessionContent(sessionId, cwd);
-            if (!content.usage) {
-                // EnterWorktree relocate (2026-08-04): jsonl read silently returned no
-                // usage (e.g. transcript moved mid-session). Surface it so token-stat
-                // fallback to per-run live usage is visible in logs, not silent.
-                getLogger().warn(`[bridge] resolveFinalUsage: no usage from jsonl sessionId=${sessionId} cwd=${cwd} agent=${agentKind}, card falls back to per-run live usage`);
-            }
-            return content.usage
-                ? {
-                    contextLength: content.usage.contextLength,
-                    contextLimit: content.usage.contextLimit,
-                    compactCount: content.usage.compactCount,
-                    compactPreContextLength: content.usage.compactPreContextLength,
-                    cacheReadTokens: content.usage.cacheReadTokens,
-                    cacheCreationTokens: content.usage.cacheCreationTokens,
-                    totalTokens: content.usage.totalTokens,
-                    inputTokens: content.usage.inputTokens,
-                    outputTokens: content.usage.outputTokens,
-                    cumulativeTotalTokens: content.usage.cumulativeTotalTokens,
-                    cumulativeInputTokens: content.usage.cumulativeInputTokens,
-                    cumulativeOutputTokens: content.usage.cumulativeOutputTokens,
-                    cumulativeCacheReadTokens: content.usage.cumulativeCacheReadTokens,
-                    cumulativeCacheCreationTokens: content.usage.cumulativeCacheCreationTokens,
-                }
-                : undefined;
-        }
-        catch (err) {
-            getLogger().warn(`[bridge] resolveFinalUsage failed sessionId=${sessionId}: ${errorMessage(err)}`);
-            return undefined;
-        }
-    }
     /**
      * Send a notification card when a run completes.
      * This card contains the session summary and a button to resume the session.
@@ -1528,8 +1494,7 @@ export class Bridge {
             return;
         // Check that the runner has runCompact method
         const runner = this.getRunner(cwd, last.agentKind);
-        if (!('runCompact' in runner) ||
-            typeof runner.runCompact !== 'function') {
+        if (!runnerHasRunCompact(runner)) {
             await this.sendResult({ text: '⚠️ 当前运行模式不支持 Compact' }, ctx);
             return;
         }
@@ -1601,9 +1566,6 @@ export class Bridge {
             replyTo: ctx.messageId,
             runId: compactRunId,
             renderOptions: {
-                showThinking: this.config.output.showThinking,
-                showToolUse: this.config.output.showToolUse,
-                showToolResult: this.config.output.showToolResult,
                 agentKind,
             },
         });
@@ -1637,10 +1599,10 @@ export class Bridge {
             // 压缩结束后从会话 jsonl 读权威统计（压缩后上下文、压缩次数、本次压缩
             // token 消耗与会话累计）。thread/compacted 通知与 jsonl 落盘几乎同时
             // （实测差 ~9ms），compactCount 未读到则短重试，仍失败优雅降级（无统计）。
-            let finalUsage = this.resolveFinalUsage(sessionId, cwd, agentKind);
+            let finalUsage = resolveFinalUsage(this.sessionReaderRegistry, this.config.defaultAgent, sessionId, cwd, agentKind);
             for (let attempt = 0; !finalUsage?.compactCount && attempt < 2; attempt++) {
                 await sleep(150);
-                finalUsage = this.resolveFinalUsage(sessionId, cwd, agentKind);
+                finalUsage = resolveFinalUsage(this.sessionReaderRegistry, this.config.defaultAgent, sessionId, cwd, agentKind);
             }
             await cardSession.finish(terminal, {
                 resultSubtype: resultSubtype ?? 'error',
@@ -1720,8 +1682,7 @@ export class Bridge {
             return;
         // Check that the runner has runCompact（codex/kimi/opencode/pi/claude 鸭子探测）。
         const runner = this.getRunner(cwd, agentKind);
-        if (!('runCompact' in runner) ||
-            typeof runner.runCompact !== 'function') {
+        if (!runnerHasRunCompact(runner)) {
             await this.sendResult({ text: '⚠️ 当前运行模式不支持 Compact' }, ctx);
             return;
         }
@@ -1861,7 +1822,4 @@ export class Bridge {
         }
     }
 }
-function errorMessage(error) {
-    return error instanceof Error ? error.message : String(error);
-}
 //# sourceMappingURL=index.js.map

package/dist/bridge/queue-manager.js

@@ -1,4 +1,5 @@
 import { getLogger } from '../logger/index.js';
+import { displayName } from '../platform/path.js';
 /**
  * Per-workspace cap on the number of tasks WAITING to execute (review P2-5).
  * Without a bound, a message flood grows `queuedTasks` / the promise chain /
@@ -377,7 +378,7 @@ export class QueueManager {
     }
     /** Build queue status card elements. */
     buildQueueStatusCardElements(workspace, actualPosition, tasksAhead, messageId, messagePreview) {
-        const workspaceName = workspace.split('/').pop() ?? workspace;
+        const workspaceName = displayName(workspace);
         const isRunning = this.isWorkspaceRunning(workspace);
         const elements = [
             {
@@ -471,7 +472,7 @@ export class QueueManager {
             getLogger().debug(`[queue-manager] no queue card to update, messageId=${messageId}`);
             return;
         }
-        const workspaceName = workspace.split('/').pop() ?? workspace;
+        const workspaceName = displayName(workspace);
         try {
             // Prefer the live preview: an edited task's messagePreview is updated in
             // place, and the passed-in snapshot may be stale if the card send was in
@@ -547,7 +548,7 @@ export class QueueManager {
             getLogger().debug(`[queue-manager] no queue card to update (cancelled), messageId=${messageId}`);
             return;
         }
-        const workspaceName = workspace.split('/').pop() ?? workspace;
+        const workspaceName = displayName(workspace);
         const card = {
             schema: '2.0',
             config: { wide_screen_mode: true },