@myskyline_ai/ccdebug 0.2.17 → 0.2.18

package/README.en.md

@@ -1,131 +1,141 @@
+
 # CCDebug - Claude Code Debugging Tool
 
-CCDebug is a debugging tool for Claude Code. It records and visualizes Claude Code execution traces, and also supports “edit & replay” for a single LLM request so you can quickly pinpoint deviations caused by prompts/context/tool calls.
+CCDebug is a debugging tool for Claude Code. It records and visualizes Claude Code execution traces, and also supports “edit & replay” for a single LLM request to help you quickly pinpoint deviations caused by prompts/context/tool calls. (This project is a derivative work based on [lemmy/claude-trace](https://github.com/badlogic/lemmy/tree/main/apps/claude-trace).)
 
-This project is a derivative work based on [lemmy/claude-trace](https://github.com/badlogic/lemmy/tree/main/apps/claude-trace).
+[>>中文 README](./README.md)
 
 ## ✨ Key Features
 
-### 📊 Timeline view for Claude Code execution
+### **1. Timeline trace view**
+- **Show by Claude Code execution steps**: distinguish different log types such as user input, LLM replies, and tool calls
+- **Step details supported**: quickly inspect the raw Claude Code log content for any step
+  
+![Timeline trace for Claude Code](./docs/img/1、日志以时间线形式展示.png)
+
+### **2. Quick log switching**
+- **One-click jump to SubAgent logs**: jump to sub-agent logs to inspect the sub-agent execution trace
+- **Quickly switch projects and sessions**: switch between different projects and sessions under `~/.claude/projects` without running the tool multiple times
+
+![Quickly switch projects and sessions](./docs/img/2、快速切换项目和会话.gif)
 
-![Timeline](./docs/img/时间线.png)
+### **3. Fast search & pinpointing**
+- **Global keyword search**: quickly locate logs containing a given keyword (searches all log files under the current session)
+- **Step overview filtering**: the overview shows the whole execution; filter timeline nodes by type
+- **Generate share links**: generate a share link for the current session for collaborative analysis
 
-- **Conversation timeline**: visualize the full conversation flow and tool call chain.
-- **Filter by step type**: filter timeline nodes by type (user message, assistant reply, tool call, etc.).
-- **Combined tool call + result**: tool input and its output are shown together for easier inspection.
-- **Session selector & sub-agent labels**: choose a main session log; sub-agent logs are shown using agent name/description.
-- **Project switching**: switch projects under `~/.claude/projects` in the Web UI to avoid running multiple servers.
+![Fast search & pinpointing](./docs/img/3、快速搜索定位.gif)
 
-### 🛠️ Step-level LLM request debugging
+### **4. Quick latency analysis**
+- **Show step start time**: each log node displays the start time of the step
+- **Show tool-call duration**: for tool calls and SubAgent steps, nodes display the duration
+- **Measure time between steps**: mark two step nodes to automatically calculate and display the elapsed time between them
 
-![LLM Request Debug](./docs/img/LLM请求调试.png)
+![Quick latency analysis](./docs/img/4、快速分析耗时.gif)
 
-![Send Modified Request](./docs/img/发送修改后的LLM请求.png)
+### **5. Step-level debugging for Claude Code**
+> Note: the debugging feature only supports Claude Code installed via the NPM script form, and does not support the native binary version.
 
-- **Track LLM requests**: record all LLM request/response logs made by Claude Code.
-- **Replay requests**: edit the request payload and resend it to validate the response repeatedly.
+![Step-level debugging for Claude Code](./docs/img/LLM请求调试.png)
+- **Track LLM requests**: record all Claude Code request logs to the LLM in detail
+- **Resend LLM requests**: edit the LLM request payload and resend it, making it easy to repeatedly validate whether the response meets expectations
 
 ## 🚀 Quick Start
 
 ### Install
 
 ```bash
-# Recommended: install from npm
+# Recommended: install globally from npm
 npm install -g @myskyline_ai/ccdebug
 
-# Or install from a local tgz artifact
+# Or: install a local/released tgz artifact
 # npm install -g /path/to/@myskyline_ai-ccdebug-x.y.z.tgz
 ```
 
 ### Basic usage
 
-#### 1) Launch Claude and record traffic
+#### 1. Launch Claude and record interactions
 
 ```bash
-# Basic usage - start Claude and record logs
+# Basic usage - start Claude and record automatically
 ccdebug
 
-# Include all requests (not only /v1/messages)
+# Include all requests (not only conversations)
 ccdebug --include-all-requests
 
-# Pass all subsequent arguments to Claude (example)
-ccdebug --run-with -p "Do the work as requested" --verbose
+# Pass subsequent args to the Claude process (example)
+ccdebug --run-with -p "Please work as requested" --verbose
 ```
 
-#### 2) Start the Web timeline server
+#### 2. Start the Web site to view the timeline trace
 
 ```bash
-# Start the Web server (default port: 3001, default project dir: current working directory)
-ccdebug --serve
+# Start the Web server for the timeline (default port: 3001; default project dir: current directory)
+ccdebug -l
 
-# Custom port
-ccdebug --serve --port 3001
+# Start on a custom port
+ccdebug -l --port 3001
 
-# Specify project directory
-ccdebug --serve --project /path/to/your/cc_workdir
+# Start with a specified project directory
+ccdebug -l --project /path/to/your/cc_workdir
 ```
 
 ### Log output directories
 
 - **Claude Code standard logs (for timeline)**: `.claude-trace/cclog/*.jsonl` (includes main logs and `agent-*.jsonl` sub-agent logs)
-- **Claude API tracing logs (for LLM debugging)**: `.claude-trace/tracelog/*.jsonl`
-- **Saved LLM request overrides (for replay)**: `.claude-trace/tracelog/llm_requests/*.json`
+- **Claude Code API tracing logs (for LLM request debugging)**: `.claude-trace/tracelog/*.jsonl`
+- **Saved LLM requests (for override/replay)**: `.claude-trace/tracelog/llm_requests/*.json`
 
 Notes:
 
-- After a `ccdebug`-launched Claude session ends, CCDebug automatically copies the corresponding Claude Code standard logs into `.claude-trace/cclog/`, and renames the API tracing log to `{sessionId}.jsonl`.
-- If you are using the **native Claude Code binary** (not the npm script version), CCDebug cannot intercept API requests, so LLM request debugging won’t work. You can still use the Web UI to view existing standard logs.
+- After you run `ccdebug` to start a Claude session, CCDebug automatically copies the corresponding Claude Code standard logs into `.claude-trace/cclog/`, and renames the API tracing log to `{sessionId}.jsonl`.
+- If you are using the **native Claude Code binary version** (not the NPM script form), CCDebug cannot intercept API requests, and LLM request debugging will be unavailable (you can still view existing Claude Code standard logs via the Web site).
 
 ## 📋 CLI Options
 
 | Option | Description |
 |------|------|
-| `--serve` | Start the Web timeline server |
-| `--log, -l` | Start the Web timeline server (`--log` without a value behaves like `--serve`; prefer `-l` to avoid confusion with `--log <name>`) |
+| `--log, -l` | Start the Web timeline server (when `--log` is used without a value, it is equivalent to `--serve`; prefer `-l` to avoid confusion with `--log <name>`) |
 | `--port <number>` | Web server port (default: 3001) |
-| `--project <path>` | Project directory |
+| `--project <path>` | Project directory path |
 | `--run-with <args>` | Pass subsequent args to the Claude process |
-| `--include-all-requests` | Capture all Claude-related requests, not only chat requests |
-| `--no-open` | Do not auto-open the generated HTML in a browser (currently only effective for `--generate-html`) |
-| `--claude-path <path>` | Custom path to the Claude binary or `cli.js` |
-| `--log <name>` | Base name for API tracing logs (affects files under `.claude-trace/tracelog/`) |
-| `--generate-html <input.jsonl> [output.html]` | Generate an HTML report from a JSONL file |
-| `--index` | Generate conversation summaries & an index under `.claude-trace/` (will call Claude and incur token usage) |
-| `--extract-token` | Extract OAuth token and exit |
-| `--version, -v` | Print version |
+| `--include-all-requests` | Include all fetch requests, not only conversations |
+| `--claude-path <path>` | Custom path to the Claude binary |
+| `--index` | Generate conversation summaries and an index for the `.claude-trace/` directory (will call Claude and incur extra token usage) |
+| `--version, -v` | Show version information |
 | `--help, -h` | Show help |
 
 ## 🏗️ Architecture
 
 ### Core components
 
-- **HTTP/API interceptor**: intercepts Node.js HTTP/HTTPS + fetch to capture Anthropic/Bedrock requests and responses.
-- **Standard log collection**: on exit, copies main and sub-agent Claude Code logs into `.claude-trace/cclog/`.
-- **Web server**: Express provides APIs for file listing, session management, project switching, and LLM request read/save/replay.
-- **Frontend**: Vue 3 + Vite + Pinia + Arco Design.
+- **HTTP/API interceptor**: based on Node.js HTTP/HTTPS + fetch interception, recording requests and responses to Anthropic/Bedrock
+- **Claude Code standard log consolidation**: on session exit, automatically copies Claude Code standard logs (main and sub-agent logs) into `.claude-trace/cclog/`
+- **Web server**: Express.js provides APIs for file listing, session management, project switching, and LLM request read/save/replay
+- **Frontend UI**: Vue 3 + Vite + Pinia + Arco Design
 
 ### Data flow
 
 ```
-HTTP request/response → interceptor → raw JSONL → processors → structured data → Web UI
+HTTP request/response → interceptor → raw data (JSONL) → data processor → structured data → Web UI
 ```
 
 ## 📁 Project Structure
 
 ```
 ccdebug/
-├── src/                     # CLI & interceptors
+├── src/                     # CLI and interceptors
 │   ├── cli.ts              # CLI entry
 │   ├── interceptor.ts      # API interception and tracelog recording
 │   ├── html-generator.ts   # HTML report generator (based on frontend)
-│   └── index-generator.ts  # conversation summaries and index
+│   └── index-generator.ts  # conversation summaries and index generation
 ├── web/                     # Web timeline site (Vite + Vue 3)
 │   ├── src/                # frontend source
 │   ├── dist/               # build output
-│   └── server/             # Express backend (required by CLI to start)
+│   └── server/             # Express backend (started by the CLI via require)
 ├── frontend/                # standalone HTML report frontend (bundle injected into HTML)
 ├── scripts/                 # packaging scripts
-└── docs/                    # docs and design notes
+└── docs/                    # documentation and design notes
 ```
 
 ## 🔧 Development
@@ -138,7 +148,7 @@ ccdebug/
 ### Local development
 
 ```bash
-# Clone
+# Clone the repo
 git clone https://github.com/ThinkingBeing/ccdebug.git
 cd ccdebug
 
@@ -148,14 +158,14 @@ npm install
 # Build
 npm run build
 
-# Dev mode (watch core code + web frontend)
+# Development mode (watch core code + frontend)
 npm run dev
 
-# Run CLI via tsx (for development/debugging)
+# Run the CLI via tsx (for development debugging)
 npx tsx src/cli.ts --help
-npx tsx src/cli.ts --serve --port 3001 --project /path/to/your/cc_workdir
+npx tsx src/cli.ts -l --port 3001 --project /path/to/your/cc_workdir
 
-# Package (artifacts will be placed under release/)
+# Package after validation (artifacts will be under release/)
 npm run package
 ```
 

package/README.md

@@ -1,23 +1,42 @@
 # CCDebug - Claude Code 调试工具
 
 
-CCDebug 是一个针对 Claude Code 的调试工具，用于记录并可视化 CC 运行轨迹，同时支持对单个 LLM 请求进行“修改-重放”，帮助你快速定位提示词/上下文/工具调用导致的偏差。本项目基于 [lemmy/claude-trace](https://github.com/badlogic/lemmy/tree/main/apps/claude-trace) 二次开发。
+CCDebug 是一个针对 Claude Code 的调试工具，用于记录并可视化 CC 运行轨迹，同时支持对单个 LLM 请求进行“修改-重放”，帮助你快速定位提示词/上下文/工具调用导致的偏差。（本项目基于 [lemmy/claude-trace](https://github.com/badlogic/lemmy/tree/main/apps/claude-trace) 二次开发）
 
-英文版：[README.en.md](./README.en.md)
+[>>English README](./README.en.md)
 
 ## ✨ 主要功能
 
-### 📊 时间线展示CC运行轨迹
-![时间线展示CC运行轨迹](./docs/img/时间线.png)
-- **对话时间线**: 直观展示完整的对话流程和工具调用链
-- **按节点类型过滤**: 按不同类型的时间线节点过滤，如用户输入、LLM回复、工具调用等
-- **工具调用和结果整合展示**: 将工具调用和调用结果整合在一个节点展示，方便查看工具调用的输入参数和输出结果
-- **会话选择与子代理识别**: 支持主日志会话选择，子代理日志下拉框显示 agent 名称/描述
-- **项目切换**: Web 站点内可切换到 `~/.claude/projects` 下的其他项目，避免重复启动多个站点
+### **1、时间线展示轨迹**
+- **按CC执行步骤显示**: 可以区分不同类型日志，用户输入、LLM回复、工具调用等
+- **支持步骤详情**: 可以快速查看某一步的原始CC日志内容
+  
+![时间线展示CC运行轨迹](./docs/img/1、日志以时间线形式展示.png)
+
+### **2、快速切换日志**
+- **一键跳转SubAgent日志**: 一键跳转到子代理日志，方便查看子代理的运行轨迹
+- **快速切换项目和会话**: 直接切换 `~/.claude/projects` 下的不同项目和会话，无需启动多次工具
+
+![快速切换项目和会话](./docs/img/2、快速切换项目和会话.gif)
+
+### **3、快速搜索定位**
+- **全局搜索关键字**: 可以快速定位到包含指定关键字的日志（搜索当前会话下所有日志文件）
+- **步骤概览过滤**: 步骤概览显示执行全貌，可按不同类型的时间线节点过滤
+- **生成分享链接**: 可以生成当前会话的分享链接，方便多人协作分析
+
+![快速搜索定位](./docs/img/3、快速搜索定位.gif)
+
+### **4、快速分析耗时**
+- **显示步骤开始时间**: 每个日志节点显示当前步骤的开始时间
+- **显示工具调用耗时**: 对于工具调用、SubAgent执行的步骤，节点将显示调用耗时
+- **统计步骤间耗时**: 用户可以标记2个步骤节点，共计自动计算并显示2个节点间耗时
+
+![快速分析耗时](./docs/img/4、快速分析耗时.gif)
+
+### **5、对CC步骤单点调试**
+> 注意：调试功能仅支持NPM脚本形式安装的 Claude Code，不支持原生二进制版本。
 
-### 🛠️ 对CC步骤单点调试
 ![对CC步骤单点调试](./docs/img/LLM请求调试.png)
-![发送修改后的LLM请求](./docs/img/发送LLM请求.png)
 - **追踪LLM请求**: 详细记录CC对LLM的所有请求日志
 - **重新发送LLM请求**: 支持修改LLM请求数据，重新发送请求，方便反复验证LLM的响应是否符合预期
 
@@ -52,13 +71,13 @@ ccdebug --run-with -p "请按要求工作" --verbose
 
 ```bash
 # 启动 Web 服务器查看时间线（默认端口 3001，默认项目目录为当前目录）
-ccdebug --serve
+ccdebug -l
 
 # 在自定义端口启动
-ccdebug --serve --port 3001
+ccdebug -l --port 3001
 
 # 指定项目目录启动
-ccdebug --serve --project /path/to/your/cc_workdir
+ccdebug -l --project /path/to/your/cc_workdir
 ```
 
 ### 日志输出目录
@@ -76,18 +95,13 @@ ccdebug --serve --project /path/to/your/cc_workdir
 
 | 选项 | 描述 |
 |------|------|
-| `--serve` | 启动 Web 时间线服务器 |
 | `--log, -l` | 启动 Web 时间线服务器（当 `--log` 不带参数时等同于 `--serve`；建议用 `-l` 避免与 `--log <name>` 混淆） |
 | `--port <number>` | 指定 Web 服务器端口（默认 3001） |
 | `--project <path>` | 指定项目目录路径 |
 | `--run-with <args>` | 将后续参数传递给 Claude 进程 |
 | `--include-all-requests` | 包含所有 fetch 请求，而不仅仅是对话 |
-| `--no-open` | 不在浏览器中自动打开生成的 HTML（当前仅对 `--generate-html` 生效） |
 | `--claude-path <path>` | 指定 Claude 二进制文件的自定义路径 |
-| `--log <name>` | 指定 API 跟踪日志基础名称（影响 `.claude-trace/tracelog/` 下文件名） |
-| `--generate-html <input.jsonl> [output.html]` | 从 JSONL 文件生成 HTML 报告 |
 | `--index` | 为 `.claude-trace/` 目录生成对话摘要和索引（会调用 Claude 产生额外 token 消耗） |
-| `--extract-token` | 提取 OAuth token 并退出 |
 | `--version, -v` | 显示版本信息 |
 | `--help, -h` | 显示帮助信息 |
 
@@ -149,7 +163,7 @@ npm run dev
 
 # 使用 tsx 直接运行 CLI（开发调试用）
 npx tsx src/cli.ts --help
-npx tsx src/cli.ts --serve --port 3001 --project /path/to/your/cc_workdir
+npx tsx src/cli.ts -l --port 3001 --project /path/to/your/cc_workdir
 
 # 功能ok后打包，包会出现在release目录
 npm run package

package/dist/cli.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"cli.d.ts","sourceRoot":"","sources":["../src/cli.ts"],"names":[],"mappings":";AAyBA,eAAO,MAAM,MAAM;;;;;;CAMT,CAAC"}
+{"version":3,"file":"cli.d.ts","sourceRoot":"","sources":["../src/cli.ts"],"names":[],"mappings":";AA2BA,eAAO,MAAM,MAAM;;;;;;CAMT,CAAC"}

package/dist/cli.js

@@ -38,7 +38,8 @@ exports.colors = void 0;
 const child_process_1 = require("child_process");
 const path = __importStar(require("path"));
 const fs = __importStar(require("fs"));
-const html_generator_1 = require("./html-generator");
+const crypto = __importStar(require("crypto"));
+const log_file_manager_1 = require("./log-file-manager");
 const os = __importStar(require("os"));
 /**
  * 获取工具版本号
@@ -320,12 +321,132 @@ function getLoaderPath() {
     }
     return loaderPath;
 }
-// Scenario 1: No args -> launch node with interceptor and absolute path to claude
-async function runClaudeWithInterception(claudeArgs = [], includeAllRequests = false, openInBrowser = false, customClaudePath, logBaseName, enableTrace = false) {
+/**
+ * 拷贝 Claude Code 日志文件到 cclog 目录
+ * @param sessionId 会话ID
+ * @param ccLogDir cclog 目录路径
+ */
+function copyCClogFile(sessionId, ccLogDir) {
+    // 检查是否启用了跟踪
+    if (!ccLogDir) {
+        return;
+    }
+    // 将当前会话对应的 cc 日志文件，拷贝到 .claude-trace/cclog 目录
+    try {
+        // 创建 LogFileManager 实例
+        const logFileManager = new log_file_manager_1.LogFileManager();
+        // 获取当前工作目录作为项目路径
+        const currentProjectPath = process.cwd();
+        // 通过 LogFileManager 解析源日志目录
+        const sourceLogDir = logFileManager.resolveLogDirectory(currentProjectPath);
+        // 构建源文件路径（假设源文件名为 sessionId.jsonl）
+        const sourceFile = path.join(sourceLogDir, `${sessionId}.jsonl`);
+        // 构建目标文件路径
+        const ccLogFile = path.join(ccLogDir, `${sessionId}.jsonl`);
+        // 检查源文件是否存在
+        if (!fs.existsSync(sourceFile)) {
+            console.log(`源CC日志文件不存在: ${sourceFile}`);
+            return;
+        }
+        // 确保目标 ccLogDir 目录存在（修复：之前未创建目录导致拷贝失败）
+        if (!fs.existsSync(ccLogDir)) {
+            fs.mkdirSync(ccLogDir, { recursive: true });
+            console.log(`已创建 cclog 目录: ${ccLogDir}`);
+        }
+        // 拷贝文件
+        fs.copyFileSync(sourceFile, ccLogFile);
+        console.log(`CC日志文件已从 ${sourceFile} 拷贝到 ${ccLogFile}`);
+        // 读取 sourceLogDir 目录下所有 agent_*.jsonl 文件，读取第一条记录的 sessionId，找到与 sessionId 变量值相同的文件，拷贝到 ccLogDir 目录
+        const files = fs.readdirSync(sourceLogDir).filter(file => file.startsWith('agent-') && file.endsWith('.jsonl'));
+        for (const file of files) {
+            const filePath = path.join(sourceLogDir, file);
+            const content = fs.readFileSync(filePath, 'utf-8');
+            const lines = content.split('\n').filter(line => line.trim());
+            if (lines.length > 0) {
+                try {
+                    const firstRecord = JSON.parse(lines[0]);
+                    const recordSessionId = firstRecord?.sessionId;
+                    if (recordSessionId === sessionId) {
+                        // 构建目标文件路径
+                        const ccAgentLogFile = path.join(ccLogDir, file);
+                        // 拷贝文件
+                        fs.copyFileSync(filePath, ccAgentLogFile);
+                        console.log(`SubAgent的CC日志文件已从 ${filePath} 拷贝到 ${ccAgentLogFile}`);
+                    }
+                }
+                catch (parseError) {
+                    // 静默处理解析错误，继续下一个文件
+                    continue;
+                }
+            }
+        }
+        // 兼容 subagents 日志的另一种存放方式：sourceLogDir/{sessionId}/subagents/ 目录
+        const subagentsDir = path.join(sourceLogDir, sessionId, 'subagents');
+        if (fs.existsSync(subagentsDir)) {
+            try {
+                const subagentFiles = fs.readdirSync(subagentsDir).filter(file => file.startsWith('agent-') && file.endsWith('.jsonl'));
+                // 只有当存在需要拷贝的文件时，才创建目标目录
+                if (subagentFiles.length > 0) {
+                    // 创建目标目录结构：ccLogDir/{sessionId}/subagents/
+                    const targetSubagentsDir = path.join(ccLogDir, sessionId, 'subagents');
+                    if (!fs.existsSync(targetSubagentsDir)) {
+                        fs.mkdirSync(targetSubagentsDir, { recursive: true });
+                    }
+                    for (const file of subagentFiles) {
+                        const sourceFilePath = path.join(subagentsDir, file);
+                        const targetFilePath = path.join(targetSubagentsDir, file);
+                        // 直接拷贝文件，因为已经在正确的 sessionId 目录下
+                        fs.copyFileSync(sourceFilePath, targetFilePath);
+                        console.log(`SubAgent的CC日志文件已从 ${sourceFilePath} 拷贝到 ${targetFilePath}`);
+                    }
+                }
+            }
+            catch (error) {
+                console.log(`处理subagents目录时出错: ${error}`);
+            }
+        }
+    }
+    catch (error) {
+        console.log(`拷贝CC日志文件时出错: ${error}`);
+    }
+}
+/**
+ * 根据 sessionId 重命名跟踪日志文件
+ * @param sessionId 会话ID
+ * @param traceLogFile 跟踪日志文件路径
+ * @returns 新的日志文件路径
+ */
+function renameTraceLogFileBySessionId(sessionId, traceLogFile) {
+    // 检查是否启用了跟踪
+    if (!traceLogFile) {
+        return null;
+    }
+    try {
+        const logDir = path.dirname(traceLogFile);
+        const newLogFile = path.join(logDir, `${sessionId}.jsonl`);
+        // 如果源文件不存在（如原生二进制模式下 interceptor 未运行，tracelog 未创建），直接跳过
+        if (!fs.existsSync(traceLogFile)) {
+            return null;
+        }
+        // 重命名文件
+        fs.renameSync(traceLogFile, newLogFile);
+        console.log(`Log file renamed from ${path.basename(traceLogFile)} to ${sessionId}.jsonl`);
+        return newLogFile;
+    }
+    catch (error) {
+        console.log(`Error renaming log file: ${error}`);
+        return traceLogFile;
+    }
+}
+// 启动claude code
+async function runClaudeWithInterception(claudeArgs = [], includeAllRequests = false, openInBrowser = false, customClaudePath, logBaseName, enableTrace = true) {
     log("启动 Claude", "blue");
     if (claudeArgs.length > 0) {
         log(`Claude 参数: ${claudeArgs.join(" ")}`, "blue");
     }
+    // 生成 UUID 格式的 sessionId
+    const sessionId = crypto.randomUUID();
+    log(`生成 SessionId: ${sessionId}`, "blue");
     const claudePath = getClaudeAbsolutePath(customClaudePath);
     log(`使用 Claude 二进制文件: ${claudePath}`, "blue");
     let child;
@@ -338,7 +459,9 @@ async function runClaudeWithInterception(claudeArgs = [], includeAllRequests = f
             log("使用 Node.js 启动 Claude（未启用跟踪）", "blue");
         }
         const loaderPath = getLoaderPath();
-        const spawnArgs = ["--require", loaderPath, claudePath, ...claudeArgs];
+        // 添加 --session-id 参数到 Claude 启动参数中
+        const claudeArgsWithSession = ["--session-id", sessionId, ...claudeArgs];
+        const spawnArgs = ["--require", loaderPath, claudePath, ...claudeArgsWithSession];
         child = (0, child_process_1.spawn)("node", spawnArgs, {
             env: {
                 ...process.env,
@@ -362,8 +485,10 @@ async function runClaudeWithInterception(claudeArgs = [], includeAllRequests = f
         console.log("");
         // 给用户一点时间阅读提示信息
         await new Promise(resolve => setTimeout(resolve, 500));
+        // 添加 --session-id 参数到 Claude 启动参数中
+        const claudeArgsWithSession = ["--session-id", sessionId, ...claudeArgs];
         // 直接启动 Claude 二进制文件，不使用代理
-        child = (0, child_process_1.spawn)(claudePath, claudeArgs, {
+        child = (0, child_process_1.spawn)(claudePath, claudeArgsWithSession, {
             env: {
                 ...process.env,
             },
@@ -418,118 +543,29 @@ async function runClaudeWithInterception(claudeArgs = [], includeAllRequests = f
         log(`意外错误: ${err.message}`, "red");
         process.exit(1);
     }
-}
-// Scenario 2: --extract-token -> launch node with token interceptor and absolute path to claude
-async function extractToken(customClaudePath) {
-    const claudePath = getClaudeAbsolutePath(customClaudePath);
-    // Log to stderr so it doesn't interfere with token output
-    console.error(`使用 Claude 二进制文件: ${claudePath}`);
-    // Create .claude-trace directory if it doesn't exist
-    const ccdebugDir = path.join(process.cwd(), ".claude-trace");
-    if (!fs.existsSync(ccdebugDir)) {
-        fs.mkdirSync(ccdebugDir, { recursive: true });
-    }
-    // Token file location
-    const tokenFile = path.join(ccdebugDir, "token.txt");
-    // Use the token extractor directly without copying
-    const tokenExtractorPath = path.join(__dirname, "token-extractor.js");
-    if (!fs.existsSync(tokenExtractorPath)) {
-        log(`未找到令牌提取器: ${tokenExtractorPath}`, "red");
-        process.exit(1);
-    }
-    const cleanup = () => {
-        try {
-            if (fs.existsSync(tokenFile))
-                fs.unlinkSync(tokenFile);
-        }
-        catch (e) {
-            // Ignore cleanup errors
-        }
-    };
-    // Launch node with token interceptor and absolute path to claude
-    const { ANTHROPIC_API_KEY, ...envWithoutApiKey } = process.env;
-    const child = (0, child_process_1.spawn)("node", ["--require", tokenExtractorPath, claudePath, "-p", "hello"], {
-        env: {
-            ...envWithoutApiKey,
-            NODE_TLS_REJECT_UNAUTHORIZED: "0",
-            CLAUDE_TRACE_TOKEN_FILE: tokenFile,
-        },
-        stdio: "inherit", // Suppress all output from Claude
-        cwd: process.cwd(),
-    });
-    // Set a timeout to avoid hanging
-    const timeout = setTimeout(() => {
-        child.kill();
-        cleanup();
-        console.error("超时: 30 秒内未找到令牌");
-        process.exit(1);
-    }, 30000);
-    // Handle child process events
-    child.on("error", (error) => {
-        clearTimeout(timeout);
-        cleanup();
-        console.error(`启动 Claude 时出错: ${error.message}`);
-        process.exit(1);
-    });
-    child.on("exit", () => {
-        clearTimeout(timeout);
-        try {
-            if (fs.existsSync(tokenFile)) {
-                const token = fs.readFileSync(tokenFile, "utf-8").trim();
-                cleanup();
-                if (token) {
-                    // Only output the token, nothing else
-                    console.log(token);
-                    process.exit(0);
-                }
-            }
-        }
-        catch (e) {
-            // File doesn't exist or read error
-        }
-        cleanup();
-        console.error("未找到授权令牌");
-        process.exit(1);
-    });
-    // Check for token file periodically
-    const checkToken = setInterval(() => {
-        try {
-            if (fs.existsSync(tokenFile)) {
-                const token = fs.readFileSync(tokenFile, "utf-8").trim();
-                if (token) {
-                    clearTimeout(timeout);
-                    clearInterval(checkToken);
-                    child.kill();
-                    cleanup();
-                    // Only output the token, nothing else
-                    console.log(token);
-                    process.exit(0);
-                }
-            }
-        }
-        catch (e) {
-            // Ignore read errors, keep trying
-        }
-    }, 500);
-}
-// Scenario 3: --generate-html input.jsonl output.html
-async function generateHTMLFromCLI(inputFile, outputFile, includeAllRequests = false, openInBrowser = false) {
+    // Claude 执行完成后，处理 CC 日志文件（不管是否启用了跟踪）
     try {
-        const htmlGenerator = new html_generator_1.HTMLGenerator();
-        const finalOutputFile = await htmlGenerator.generateHTMLFromJSONL(inputFile, outputFile, includeAllRequests);
-        if (openInBrowser) {
-            (0, child_process_1.spawn)("open", [finalOutputFile], { detached: true, stdio: "ignore" }).unref();
-            log(`正在浏览器中打开 ${finalOutputFile}`, "green");
-        }
-        process.exit(0);
+        console.log("\nClaude 执行完成，处理 CC 日志文件...");
+        // 构建日志文件路径
+        const traceHomeDir = ".claude-trace";
+        const traceLogDir = path.join(traceHomeDir, 'tracelog');
+        const ccLogDir = path.join(traceHomeDir, 'cclog');
+        // 生成日志文件名
+        const fileBaseName = logBaseName || `log-${new Date().toISOString().replace(/[:.]/g, "-").replace("T", "-").slice(0, -5)}`;
+        const traceLogFile = enableTrace ? path.join(traceLogDir, `${fileBaseName}.jsonl`) : null;
+        // 直接使用之前生成的 sessionId
+        console.log(`使用 SessionId: ${sessionId} 处理日志文件`);
+        // 将当前会话对应的 cc 日志文件，拷贝到 .claude-trace/cclog 目录
+        copyCClogFile(sessionId, ccLogDir);
+        // 根据 sessionId 重命名跟踪日志文件
+        renameTraceLogFileBySessionId(sessionId, traceLogFile);
     }
     catch (error) {
-        const err = error;
-        log(`错误: ${err.message}`, "red");
-        process.exit(1);
+        // 静默处理错误，不影响主流程
+        console.log(`处理CC日志时出错: ${error}`);
     }
 }
-// Scenario 5: --serve, --log, -l
+// 启动web工具
 async function startWebServer(port, projectDir) {
     try {
         // 使用 require 导入 web server 模块
@@ -561,20 +597,6 @@ async function startWebServer(port, projectDir) {
         process.exit(1);
     }
 }
-// Scenario 4: --index
-async function generateIndex() {
-    try {
-        const { IndexGenerator } = await Promise.resolve().then(() => __importStar(require("./index-generator")));
-        const indexGenerator = new IndexGenerator();
-        await indexGenerator.generateIndex();
-        process.exit(0);
-    }
-    catch (error) {
-        const err = error;
-        log(`错误: ${err.message}`, "red");
-        process.exit(1);
-    }
-}
 // Main entry point
 async function main() {
     const args = process.argv.slice(2);
@@ -601,7 +623,7 @@ async function main() {
         process.exit(0);
     }
     // Check for trace flag
-    const enableTrace = claudeTraceArgs.includes("--trace");
+    const enableTrace = !claudeTraceArgs.includes("--notrace");
     // Check for include all requests flag
     const includeAllRequests = claudeTraceArgs.includes("--include-all-requests");
     // Check for no-open flag (inverted logic - open by default)
@@ -633,37 +655,6 @@ async function main() {
     if (projectIndex !== -1 && claudeTraceArgs[projectIndex + 1]) {
         serveProjectDir = claudeTraceArgs[projectIndex + 1];
     }
-    // Scenario 2: --extract-token
-    if (claudeTraceArgs.includes("--extract-token")) {
-        await extractToken(customClaudePath);
-        return;
-    }
-    // Scenario 3: --generate-html input.jsonl output.html
-    if (claudeTraceArgs.includes("--generate-html")) {
-        const flagIndex = claudeTraceArgs.indexOf("--generate-html");
-        const inputFile = claudeTraceArgs[flagIndex + 1];
-        // Find is next argument that's not a flag as the output file
-        let outputFile;
-        for (let i = flagIndex + 2; i < claudeTraceArgs.length; i++) {
-            const arg = claudeTraceArgs[i];
-            if (!arg.startsWith("--")) {
-                outputFile = arg;
-                break;
-            }
-        }
-        if (!inputFile) {
-            log(`--generate-html 缺少输入文件`, "red");
-            log(`用法: ccdebug --generate-html input.jsonl [output.html]`, "yellow");
-            process.exit(1);
-        }
-        await generateHTMLFromCLI(inputFile, outputFile, includeAllRequests, openInBrowser);
-        return;
-    }
-    // Scenario 4: --index
-    if (claudeTraceArgs.includes("--index")) {
-        await generateIndex();
-        return;
-    }
     // Scenario 5: --serve, --log, -l
     const hasServeFlag = claudeTraceArgs.includes("--serve");
     const hasLogFlag = claudeTraceArgs.includes("--log") || claudeTraceArgs.includes("-l");

package/dist/interceptor.d.ts

@@ -31,9 +31,6 @@ export declare class ClaudeTrafficLogger {
     private writePairToLog;
     private generateHTML;
     cleanup(): void;
-    private getSessionIdFromLog;
-    private copyCClogFile;
-    private renameTraceLogFileBySessionId;
     getStats(): {
         totalPairs: number;
         pendingRequests: number;

package/dist/interceptor.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"interceptor.d.ts","sourceRoot":"","sources":["../src/interceptor.ts"],"names":[],"mappings":"AAQA,MAAM,WAAW,iBAAiB;IACjC,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,kBAAkB,CAAC,EAAE,OAAO,CAAC;IAC7B,QAAQ,CAAC,EAAE,OAAO,GAAG,MAAM,GAAG,MAAM,GAAG,OAAO,CAAC;CAC/C;AAED,qBAAa,mBAAmB;IAC/B,OAAO,CAAC,YAAY,CAAS;IAC7B,OAAO,CAAC,WAAW,CAAS;IAC5B,OAAO,CAAC,YAAY,CAAS;IAC7B,OAAO,CAAC,QAAQ,CAAS;IACzB,OAAO,CAAC,SAAS,CAAS;IAC1B,OAAO,CAAC,QAAQ,CAAS;IACzB,OAAO,CAAC,eAAe,CAA+B;IACtD,OAAO,CAAC,KAAK,CAAiB;IAC9B,OAAO,CAAC,MAAM,CAAoB;IAClC,OAAO,CAAC,aAAa,CAAgB;gBAEzB,MAAM,GAAE,iBAAsB;IAmE1C,OAAO,CAAC,WAAW;IAqBnB,OAAO,CAAC,iBAAiB;IAIzB,OAAO,CAAC,sBAAsB;YAgChB,aAAa;YAKb,gBAAgB;YAsBhB,iBAAiB;IAwBxB,aAAa,IAAI,IAAI;IAKrB,eAAe,IAAI,IAAI;IA+FvB,kBAAkB,IAAI,IAAI;IAmDjC,OAAO,CAAC,oBAAoB;IAuE5B,OAAO,CAAC,mBAAmB;YAab,2BAA2B;YAiB3B,cAAc;YAcd,YAAY;IAmBnB,OAAO,IAAI,IAAI;IAiDtB,OAAO,CAAC,mBAAmB;IA+C3B,OAAO,CAAC,aAAa;IAoFrB,OAAO,CAAC,6BAA6B;IAuB9B,QAAQ;;;;;;CAQf;AAQD,wBAAgB,qBAAqB,CAAC,MAAM,CAAC,EAAE,iBAAiB,GAAG,mBAAmB,CA8BrF;AAED,wBAAgB,SAAS,IAAI,mBAAmB,GAAG,IAAI,CAEtD"}
+{"version":3,"file":"interceptor.d.ts","sourceRoot":"","sources":["../src/interceptor.ts"],"names":[],"mappings":"AAOA,MAAM,WAAW,iBAAiB;IACjC,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,kBAAkB,CAAC,EAAE,OAAO,CAAC;IAC7B,QAAQ,CAAC,EAAE,OAAO,GAAG,MAAM,GAAG,MAAM,GAAG,OAAO,CAAC;CAC/C;AAED,qBAAa,mBAAmB;IAC/B,OAAO,CAAC,YAAY,CAAS;IAC7B,OAAO,CAAC,WAAW,CAAS;IAC5B,OAAO,CAAC,YAAY,CAAS;IAC7B,OAAO,CAAC,QAAQ,CAAS;IACzB,OAAO,CAAC,SAAS,CAAS;IAC1B,OAAO,CAAC,QAAQ,CAAS;IACzB,OAAO,CAAC,eAAe,CAA+B;IACtD,OAAO,CAAC,KAAK,CAAiB;IAC9B,OAAO,CAAC,MAAM,CAAoB;IAClC,OAAO,CAAC,aAAa,CAAgB;gBAEzB,MAAM,GAAE,iBAAsB;IAmE1C,OAAO,CAAC,WAAW;IAqBnB,OAAO,CAAC,iBAAiB;IAIzB,OAAO,CAAC,sBAAsB;YAgChB,aAAa;YAKb,gBAAgB;YAsBhB,iBAAiB;IAwBxB,aAAa,IAAI,IAAI;IAKrB,eAAe,IAAI,IAAI;IA+FvB,kBAAkB,IAAI,IAAI;IAmDjC,OAAO,CAAC,oBAAoB;IAuE5B,OAAO,CAAC,mBAAmB;YAab,2BAA2B;YAiB3B,cAAc;YAcd,YAAY;IAmBnB,OAAO,IAAI,IAAI;IAuCf,QAAQ;;;;;;CAQf;AAQD,wBAAgB,qBAAqB,CAAC,MAAM,CAAC,EAAE,iBAAiB,GAAG,mBAAmB,CA8BrF;AAED,wBAAgB,SAAS,IAAI,mBAAmB,GAAG,IAAI,CAEtD"}

package/dist/interceptor.js

@@ -9,7 +9,6 @@ exports.getLogger = getLogger;
 const fs_1 = __importDefault(require("fs"));
 const path_1 = __importDefault(require("path"));
 const html_generator_1 = require("./html-generator");
-const log_file_manager_1 = require("./log-file-manager");
 class ClaudeTrafficLogger {
     constructor(config = {}) {
         this.pendingRequests = new Map();
@@ -433,14 +432,6 @@ class ClaudeTrafficLogger {
         }
         this.pendingRequests.clear();
         console.log(`Cleanup complete. Logged ${this.pairs.length} pairs`);
-        //获取cc会话的sessionid
-        let sessionId = this.getSessionIdFromLog();
-        if (sessionId != '') {
-            //将当前会话对应的cc日志文件，拷贝到.claude-trace/cclog目录
-            this.copyCClogFile(sessionId);
-            // Rename log file based on sessionid from first record
-            this.renameTraceLogFileBySessionId(sessionId);
-        }
         // Open browser if requested
         // const shouldOpenBrowser = process.env.CLAUDE_TRACE_OPEN_BROWSER === "true";
         // if (shouldOpenBrowser && fs.existsSync(this.htmlFile)) {
@@ -452,133 +443,6 @@ class ClaudeTrafficLogger {
         // 	}
         // }
     }
-    getSessionIdFromLog() {
-        // 检查是否启用了跟踪
-        if (!this.traceLogFile) {
-            return '';
-        }
-        // Check if log file exists
-        if (!fs_1.default.existsSync(this.traceLogFile)) {
-            console.log("获取sessionId错误：Log file does not exist");
-            return '';
-        }
-        // Read the first line of the JSONL file
-        const fileContent = fs_1.default.readFileSync(this.traceLogFile, 'utf-8');
-        const lines = fileContent.split('\n').filter(line => line.trim());
-        if (lines.length === 0) {
-            console.log("获取sessionId错误：Log file is empty");
-            return '';
-        }
-        // 循环读取日志，直到找到user_id为止
-        let userId = null;
-        for (const line of lines) {
-            const record = JSON.parse(line);
-            userId = record?.request?.body?.metadata?.user_id;
-            if (userId) {
-                break;
-            }
-        }
-        if (!userId) {
-            console.log("获取sessionId错误：No user_id found in any record");
-            return '';
-        }
-        // Extract sessionid from user_id (format: xxxx_session_{sessionid})
-        const sessionMatch = userId.match(/_session_([^_]+)$/);
-        if (!sessionMatch || !sessionMatch[1]) {
-            console.log(`获取sessionId错误：No sessionid found in user_id: ${userId}`);
-            return '';
-        }
-        return sessionMatch[1];
-    }
-    copyCClogFile(sessionId) {
-        // 检查是否启用了跟踪
-        if (!this.ccLogDir) {
-            return;
-        }
-        //将当前会话对应的cc日志文件，拷贝到.claude-trace/cclog目录
-        try {
-            // 创建LogFileManager实例
-            const logFileManager = new log_file_manager_1.LogFileManager();
-            // 获取当前工作目录作为项目路径
-            const currentProjectPath = process.cwd();
-            // 通过LogFileManager解析源日志目录
-            const sourceLogDir = logFileManager.resolveLogDirectory(currentProjectPath);
-            // 构建源文件路径（假设源文件名为sessionId.jsonl）
-            const sourceFile = path_1.default.join(sourceLogDir, `${sessionId}.jsonl`);
-            // 构建目标文件路径
-            this.ccLogFile = path_1.default.join(this.ccLogDir, `${sessionId}.jsonl`);
-            // 检查源文件是否存在
-            if (!fs_1.default.existsSync(sourceFile)) {
-                console.log(`源CC日志文件不存在: ${sourceFile}`);
-                return;
-            }
-            // 拷贝文件
-            fs_1.default.copyFileSync(sourceFile, this.ccLogFile);
-            console.log(`CC日志文件已从 ${sourceFile} 拷贝到 ${this.ccLogFile}`);
-            // 读取sourceLogDir目录下所有agent_*.jsonl文件，读取第一条记录的sessionId，找到与sessionId变量值相同的文件，拷贝到ccLogDir目录
-            const files = fs_1.default.readdirSync(sourceLogDir).filter(file => file.startsWith('agent-') && file.endsWith('.jsonl'));
-            for (const file of files) {
-                const filePath = path_1.default.join(sourceLogDir, file);
-                const content = fs_1.default.readFileSync(filePath, 'utf-8');
-                const lines = content.split('\n').filter(line => line.trim());
-                if (lines.length > 0) {
-                    try {
-                        const firstRecord = JSON.parse(lines[0]);
-                        const recordSessionId = firstRecord?.sessionId;
-                        if (recordSessionId === sessionId) {
-                            // 构建目标文件路径
-                            const ccAgentLogFile = path_1.default.join(this.ccLogDir, file);
-                            // 拷贝文件
-                            fs_1.default.copyFileSync(filePath, ccAgentLogFile);
-                            console.log(`SubAgent的CC日志文件已从 ${filePath} 拷贝到 ${ccAgentLogFile}`);
-                        }
-                    }
-                    catch (parseError) {
-                        // 静默处理解析错误，继续下一个文件
-                        continue;
-                    }
-                }
-            }
-            // 兼容subagents日志的另一种存放方式：sourceLogDir/{sessionId}/subagents/目录
-            const subagentsDir = path_1.default.join(sourceLogDir, sessionId, 'subagents');
-            if (fs_1.default.existsSync(subagentsDir)) {
-                try {
-                    const subagentFiles = fs_1.default.readdirSync(subagentsDir).filter(file => file.startsWith('agent-') && file.endsWith('.jsonl'));
-                    for (const file of subagentFiles) {
-                        const sourceFilePath = path_1.default.join(subagentsDir, file);
-                        const targetFilePath = path_1.default.join(this.ccLogDir, file);
-                        // 直接拷贝文件，因为已经在正确的sessionId目录下
-                        fs_1.default.copyFileSync(sourceFilePath, targetFilePath);
-                        console.log(`SubAgent的CC日志文件已从 ${sourceFilePath} 拷贝到 ${targetFilePath}`);
-                    }
-                }
-                catch (error) {
-                    console.log(`处理subagents目录时出错: ${error}`);
-                }
-            }
-        }
-        catch (error) {
-            console.log(`拷贝CC日志文件时出错: ${error}`);
-        }
-    }
-    renameTraceLogFileBySessionId(sessionId) {
-        // 检查是否启用了跟踪
-        if (!this.traceLogFile) {
-            return;
-        }
-        try {
-            const logDir = path_1.default.dirname(this.traceLogFile);
-            const newLogFile = path_1.default.join(logDir, `${sessionId}.jsonl`);
-            // Rename the file
-            fs_1.default.renameSync(this.traceLogFile, newLogFile);
-            console.log(`Log file renamed from ${path_1.default.basename(this.traceLogFile)} to ${sessionId}.jsonl`);
-            // Update the logFile path for future reference
-            this.traceLogFile = newLogFile;
-        }
-        catch (error) {
-            console.log(`Error renaming log file: ${error}`);
-        }
-    }
     getStats() {
         return {
             totalPairs: this.pairs.length,

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@myskyline_ai/ccdebug",
-	"version": "0.2.17",
+	"version": "0.2.18",
 	"description": "跟踪记录CC请求，可针对特定请求反复调试验证，以更直观的时间线形式展示CC日志",
 	"publishConfig": {
 		"access": "public"