merge upstream main (v0.1.26): resolve conflicts in .gitignore, App.tsx, PromptInput.tsx

Author: dengmik-commits

.deepcode/AGENTS.md

@@ -9,24 +9,21 @@ src/
 ├── settings.ts          # Settings resolution from ~/.deepcode/settings.json
 ├── prompt.ts            # System prompt builder, tool definitions, built-in skills
 ├── common/
+│   ├── bash-timeout.ts  # Bash command timeout enforcement
+│   ├── debug-logger.ts  # Debug logging for OpenAI API calls
+│   ├── error-logger.ts  # API error logging
+│   ├── file-history.ts  # GitFileHistory — checkpoint/undo via Git branches
+│   ├── file-utils.ts    # File read/write with encoding and diff preview
 │   ├── model-capabilities.ts # Model detection and thinking-mode defaults
+│   ├── notify.ts        # Desktop notification after LLM turn completion
+│   ├── openai-client.ts # OpenAI client singleton with keep-alive agent
 │   ├── openai-thinking.ts    # OpenAI thinking request options builder
-│   ├── file-utils.ts    # File read/write with encoding and diff preview
+│   ├── permissions.ts   # Permission scoping, decisions, and tool-call tracking
+│   ├── process-tree.ts  # Process tree construction and orphan detection
 │   ├── shell-utils.ts   # Shell path resolution (Git Bash, zsh, bash)
 │   ├── state.ts         # In-memory file state and snippet tracking
-│   ├── runtime.ts       # Tool validation runtime helpers
-│   ├── notify.ts        # Desktop notification after LLM turn completion
-│   ├── debug-logger.ts  # Debug logging for OpenAI API calls
-│   └── error-logger.ts  # API error logging
-├── ui/
-│   ├── App.tsx          # Root Ink component — state, routing, session orchestration
-│   ├── PromptInput.tsx  # Multi-line input with file mentions (@), slash commands, image paste, skills
-│   ├── MessageView.tsx  # Renders assistant/tool messages with markdown
-│   ├── McpStatusList.tsx # MCP server connection status and available tools
-│   ├── ProcessStdoutView.tsx # Ctrl+O fullscreen overlay for live process stdout
-│   ├── UpdatePrompt.tsx  # UpdatePlan task list progress display
-│   ├── fileMentions.ts  # @-mention file scanning, filtering, and insertion
-│   └── ...
+│   ├── update-check.ts  # Latest-version check against npm registry
+│   └── validate.ts      # Tool validation runtime helpers (was runtime.ts)
 ├── mcp/
 │   ├── mcp-client.ts    # MCP client — JSON-RPC communication with MCP servers
 │   └── mcp-manager.ts   # MCP manager — lifecycle, tool registration, execution, status
@@ -39,12 +36,21 @@ src/
 │   ├── update-plan-handler.ts # Updates the task plan progress display
 │   ├── web-search-handler.ts  # Web search via natural language queries
 │   └── ask-user-question-handler.ts # Interactive user prompts with options
+├── ui/
+│   ├── components/      # Reusable Ink components (MessageView, DropdownMenu, ModelsDropdown, SkillsDropdown, etc.)
+│   ├── contexts/        # React contexts (AppContext, RawModeContext)
+│   ├── hooks/           # Custom hooks (cursor, useHistoryNavigation, usePasteHandling, useTerminalInput)
+│   ├── views/           # Top-level screen components (App, PromptInput, SessionList, PermissionPrompt, ProcessStdoutView, etc.)
+│   ├── core/            # Core UI logic (file-mentions, slash-commands, prompt-buffer, thinking-state, etc.)
+│   ├── utils/           # Shared utility helpers
+│   ├── index.ts         # UI module barrel exports
+│   └── constants.ts     # UI-wide constants
 ├── tests/               # One *.test.ts per source module, plus run-tests.mjs
 templates/
 ├── tools/               # Tool descriptions fed to the LLM
 ├── skills/              # Built-in skill definitions (agent-drift-guard, plan-and-execute)
-├── prompts/             # EJS templates (e.g., init_command.md.ejs)
-docs/                    # User-facing documentation (configuration, MCP, skills)
+└── prompts/             # EJS templates (e.g., init_command.md.ejs)
+docs/                    # User-facing documentation (configuration, MCP, notify, permissions)
 dist/                    # Bundled CLI output (gitignored)
 ```
 
@@ -78,13 +84,13 @@ Run the CLI locally for manual testing: `node dist/cli.js` (after `npm run bundl
 
 **Formatting/Linting**: Prettier + ESLint (typescript-eslint, react-hooks). Run `npm run check` before pushing. On commit, Husky + lint-staged auto-formats staged `*.{ts,tsx,js,mjs,cjs,ejs,jsx}` and `*.json` files.
 
-**File naming**: `kebab-case.ts` for modules, `kebab-case.tsx` for React/Ink components. Test files: `*.test.ts`.
+**File naming**: `kebab-case.ts` for modules, `kebab-case.tsx` for React/Ink components. Test files: `*.test.ts` (always kebab-case).
 
 ## Testing Guidelines
 
 - **Framework**: Node.js native test runner (`node:test`) with `tsx` for TypeScript
 - **Assertions**: `node:assert/strict`
-- **Coverage**: Target meaningful unit tests for core logic (session management, tool handlers, settings resolution, prompt buffer). Test files are in `src/tests/` matching the source module name.
+- **Coverage**: Target meaningful unit tests for core logic (session management, tool handlers, settings resolution, prompt buffer, permissions, MCP client). Test files are in `src/tests/` matching the source module name.
 - **Test naming**: `describe`/`test` blocks with descriptive names. Example: `test("SessionManager preserves structured system content when building OpenAI messages", ...)`
 - **Relaxed lint rules**: Test files allow `any` and unused vars.
 - Run all tests with `npm test` before submitting a PR. A cross-platform test runner is available at `src/tests/run-tests.mjs`.
@@ -109,13 +115,17 @@ Run the CLI locally for manual testing: `node dist/cli.js` (after `npm run bundl
 
 ## Architecture Overview
 
-The CLI (`@vegamo/deepcode-cli`) renders a terminal UI using [Ink](https://github.com/vadimdemedes/ink) (React for terminals). `SessionManager` drives the LLM interaction loop: it builds system prompts, sends user messages with optional skills/images, streams responses, executes tool calls via `ToolExecutor`, and compacts context when token thresholds are exceeded (512K for DeepSeek V4 models, 128K for others).
+The CLI (`@vegamo/deepcode-cli`) renders a terminal UI using [Ink](https://github.com/vadimdemedes/ink) (React for terminals). `SessionManager` drives the LLM interaction loop: it builds system prompts, sends user messages with optional skills/images, streams responses, executes tool calls via `ToolExecutor`, and compacts context when token thresholds are exceeded (512K for DeepSeek V4 models, 128K for others). OpenAI client connectivity is managed by `createOpenAIClient()` in `src/common/openai-client.ts`, which caches the client singleton and applies a 180-second keep-alive timeout.
 
 Seven built-in tools are available to the LLM: `bash`, `read`, `write`, `edit`, `AskUserQuestion`, `UpdatePlan`, and `WebSearch`. Tool definitions are registered in `src/tools/executor.ts` and described to the LLM via `src/prompt.ts` and `templates/tools/`. The `UpdatePlan` tool enables the LLM to display and update a structured task list in the terminal.
 
+A **permission system** (`src/common/permissions.ts`) controls tool execution scopes (read/write/delete/network/git-log, etc.) with configurable allow/deny/ask decisions. The `PermissionPrompt` view presents interactive prompts when a tool requires user approval.
+
+A **file history system** (`src/common/file-history.ts`) provides undo/checkpoint support by creating lightweight Git branches per session. The `GitFileHistory` class manages blob storage and branch references via a `.deepcode-file-history.json` manifest.
+
 **Slash commands**: `/model`, `/new`, `/init`, `/resume`, `/continue`, `/mcp`, `/exit`, plus dynamic `/skill-name` for each loaded skill.
 
-**Key UI features**: `@` file mentions in the prompt input (scans project files), `Ctrl+O` to view live process stdout in fullscreen, `Ctrl+V` to paste images, MCP server status display.
+**Key UI features**: `@` file mentions in the prompt input (scans project files), `Ctrl+O` to view live process stdout in fullscreen, `Ctrl+V` to paste images, MCP server status display, undo selector, and permission prompts.
 
 **CLI flags**: `-p <prompt>` / `--prompt` to auto-submit a prompt on launch, `-v` / `--version`, `-h` / `--help`.
 

.gitignore

@@ -7,3 +7,4 @@ dist/
 *.log
 .claude/
 .mcp.json
+.deepcode/settings.json

README-en.md

@@ -137,6 +137,10 @@ When the AI assistant completes a task, Deep Code can automatically execute a no
 
 For detailed configuration instructions, see: [docs/notify_en.md](docs/notify_en.md)
 
+### Does Deep Code only support YOLO mode?
+
+No. Deep Code has a built-in fine-grained permission control mechanism that lets you confirm operations before the AI assistant executes shell commands, reads/writes files, accesses the network, and more. You can configure each permission scope's policy — always allow, always ask, or deny — via the `permissions` field in `settings.json`. See [docs/permission.md](docs/permission.md) for details.
+
 ## Contributing
 
 Contributions are welcome! Here's how to get started:

README-zh_CN.md

@@ -122,6 +122,10 @@ Deep Code 支持 MCP（Model Context Protocol），可以连接 GitHub、浏览
 
 详细配置指南：[docs/notify.md](docs/notify.md)
 
+### Deep Code 只支持 YOLO 模式吗？
+
+不是。Deep Code 内置了细粒度的权限控制机制，支持在 AI 助手执行 Shell 命令、读写文件、访问网络等操作前进行确认。你可以通过 `settings.json` 中的 `permissions` 字段按需配置每种权限范围的策略：始终允许、始终询问、或直接拒绝。详见 [docs/permission.md](docs/permission.md)。
+
 ### 是否支持 Coding Plan？
 
 支持。只要把 `~/.deepcode/settings.json` 的 `env.BASE_URL` 配置为 OpenAI 兼容的接口地址就行。以火山方舟的 Coding Plan 为例：
@@ -136,6 +140,7 @@ Deep Code 支持 MCP（Model Context Protocol），可以连接 GitHub、浏览
   "thinkingEnabled": true
 }
 ```
+
 ## 贡献
 
 欢迎贡献代码！以下是参与方式：

README.md

@@ -122,6 +122,10 @@ Deep Code 支持 MCP（Model Context Protocol），可以连接 GitHub、浏览
 
 详细配置指南：[docs/notify.md](docs/notify.md)
 
+### Deep Code 只支持 YOLO 模式吗？
+
+不是。Deep Code 内置了细粒度的权限控制机制，支持在 AI 助手执行 Shell 命令、读写文件、访问网络等操作前进行确认。你可以通过 `settings.json` 中的 `permissions` 字段按需配置每种权限范围的策略：始终允许、始终询问、或直接拒绝。详见 [docs/permission.md](docs/permission.md)。
+
 ### 是否支持 Coding Plan？
 
 支持。只要把 `~/.deepcode/settings.json` 的 `env.BASE_URL` 配置为 OpenAI 兼容的接口地址就行。以火山方舟的 Coding Plan 为例：
@@ -136,6 +140,7 @@ Deep Code 支持 MCP（Model Context Protocol），可以连接 GitHub、浏览
   "thinkingEnabled": true
 }
 ```
+
 ## 贡献
 
 欢迎贡献代码！以下是参与方式：

docs/configuration.md

@@ -31,6 +31,7 @@ Deep Code 使用 `settings.json` 设置文件进行持久化配置，支持两
 | `thinkingEnabled`    | boolean   | 是否启用思考模式（DeepSeek V4 系列默认启用）                         |
 | `reasoningEffort`    | string    | 推理强度，可选 `"high"` 或 `"max"`（默认 `"max"`）                  |
 | `debugLogEnabled`    | boolean   | 是否启用调试日志输出（默认 `false`）                                 |
+| `telemetryEnabled`   | boolean   | 是否启用匿名使用数据上报（默认 `true`）                              |
 | `notify`             | string    | 任务完成通知脚本的完整路径（如 Slack 通知脚本）                      |
 | `webSearchTool`      | string    | 自定义联网搜索脚本的完整路径                                         |
 | `mcpServers`         | object    | MCP 服务器配置（键为服务名，值为 McpServerConfig 对象）              |
@@ -45,6 +46,7 @@ Deep Code 使用 `settings.json` 设置文件进行持久化配置，支持两
 | `THINKING_ENABLED`  | string | 是否启用思考模式                                         |
 | `REASONING_EFFORT`  | string | 推理强度                                                |
 | `DEBUG_LOG_ENABLED`  | string | 是否启用调试日志输出                                     |
+| `TELEMETRY_ENABLED`  | string | 是否启用匿名使用数据上报                                   |
 | `<其他任意KEY>` | string | 自定义环境变量 |
 
 #### `thinkingEnabled` — 思考模式
@@ -130,6 +132,16 @@ MCP（Model Context Protocol）服务器配置。值是键值对，键为服务
 
 设为 `true` 可让程序输出详细的调试日志（默认 `false`），用于排查 API 调用和工具执行的问题。
 
+#### `telemetryEnabled` — 匿名使用数据上报
+
+设为 `false` 可关闭匿名使用数据上报（默认 `true`）。上报仅包含匿名的机器标识，不包含对话内容、代码或 API 密钥。
+
+也可以通过环境变量关闭：
+
+```bash
+DEEPCODE_TELEMETRY_ENABLED=0 deepcode
+```
+
 ## 环境变量优先级
 
 环境变量是配置应用程序的常用方式，尤其适用于敏感信息（如 api-key）或可能在不同环境之间更改的设置。

docs/configuration_en.md

@@ -31,6 +31,7 @@ The following are all the top-level fields supported in `settings.json`, along w
 | `thinkingEnabled`  | boolean | Whether to enable thinking mode (enabled by default for DeepSeek V4 series)|
 | `reasoningEffort`  | string  | Reasoning intensity, either `"high"` or `"max"` (default `"max"`)          |
 | `debugLogEnabled`  | boolean | Enable debug log output (default `false`)                                   |
+| `telemetryEnabled` | boolean | Enable anonymous usage reporting (default `true`)                           |
 | `notify`           | string  | Full path to a task-completion notification script (e.g., Slack notification script) |
 | `webSearchTool`    | string  | Full path to a custom web search script                                     |
 | `mcpServers`       | object  | MCP server configurations (keys are service names, values are McpServerConfig objects) |
@@ -45,6 +46,7 @@ The following are all the top-level fields supported in `settings.json`, along w
 | `THINKING_ENABLED`| string | Enable thinking mode                                            |
 | `REASONING_EFFORT`| string | Reasoning intensity                                             |
 | `DEBUG_LOG_ENABLED`| string| Enable debug log output                                         |
+| `TELEMETRY_ENABLED`| string| Enable anonymous usage reporting                                |
 | `<any other KEY>` | string | Custom environment variable                                     |
 
 #### `thinkingEnabled` — Thinking Mode
@@ -129,6 +131,16 @@ For detailed MCP usage instructions, refer to [mcp.md](mcp.md).
 
 Set to `true` to enable detailed debug logging (default `false`), useful for troubleshooting API calls and tool execution.
 
+#### `telemetryEnabled` — Anonymous Usage Reporting
+
+Set to `false` to disable anonymous usage reporting (default `true`). The report only includes an anonymous machine identifier and does not contain conversation content, code, or API keys.
+
+You can also disable it via environment variable:
+
+```bash
+DEEPCODE_TELEMETRY_ENABLED=0 deepcode
+```
+
 ## Environment Variable Priority
 
 Environment variables are a common way to configure applications, especially for sensitive information (such as api-key) or settings that may change between environments.

docs/permission.md

@@ -0,0 +1,101 @@
+# Deep Code 权限机制
+
+Deep Code 内置了一套细粒度的权限控制机制，在 AI 助手执行工具调用（如执行 Shell 命令、读写文件、访问网络等）前，根据用户配置的策略决定是自动放行、直接拒绝、还是弹出交互式确认。
+
+## 概述
+
+每次 AI 助手调用工具时，系统会自动分析该操作涉及的**权限范围（Permission Scope）**，然后根据 `settings.json` 中的权限配置做出决策。对于需要用户确认的操作，会在终端中弹出交互式选择界面，用户可以选择：
+
+- **Yes** — 仅本次放行
+- **Yes, and always allow** — 本次放行，并将该权限范围写入项目配置文件，后续同类操作不再询问
+- **No** — 拒绝本次操作
+
+## 权限范围
+
+Deep Code 定义了以下 10 种权限范围，覆盖了工具调用的各类风险场景：
+
+| 权限范围 | 说明 |
+| -------- | ---- |
+| `read-in-cwd` | 读取当前工作区内的文件 |
+| `read-out-cwd` | 读取当前工作区外的文件 |
+| `write-in-cwd` | 在当前工作区内创建或覆写文件 |
+| `write-out-cwd` | 在当前工作区外创建或覆写文件 |
+| `delete-in-cwd` | 删除当前工作区内的文件 |
+| `delete-out-cwd` | 删除当前工作区外的文件 |
+| `query-git-log` | 查询 Git 历史（如 `git log`、`git show`、`git blame`） |
+| `mutate-git-log` | 修改 Git 历史（如 `git commit`、`git rebase`、`git tag`） |
+| `network` | 访问网络（如 `curl`、`npm install` 等联网操作） |
+| `mcp` | 调用 MCP 外部工具 |
+
+此外还有一个特殊的 `unknown` 范围，当 LLM 无法准确分类命令的副作用时使用，**`unknown` 总是触发询问**。
+
+## 权限配置
+
+在 `~/.deepcode/settings.json`（用户级）或 `.deepcode/settings.json`（项目级）中通过 `permissions` 字段配置：
+
+```json
+{
+  "permissions": {
+    "allow": [],
+    "deny": [],
+    "ask": [],
+    "defaultMode": "allowAll"
+  }
+}
+```
+
+### 配置项说明
+
+| 字段 | 类型 | 说明 |
+| ---- | ---- | ---- |
+| `allow` | `string[]` | 始终自动放行的权限范围列表 |
+| `deny` | `string[]` | 始终自动拒绝的权限范围列表 |
+| `ask` | `string[]` | 始终弹出询问的权限范围列表 |
+| `defaultMode` | `"allowAll"` \| `"askAll"` | 未在 `allow`/`deny`/`ask` 中明确列出的权限范围的默认处理方式。默认为 `"allowAll"` |
+
+### 优先级规则
+
+当一个工具调用涉及多个权限范围时，决策按以下优先级进行：
+
+1. 若任一范围命中 `deny` → **拒绝**
+2. 若任一范围命中 `ask` → **询问**
+3. 若所有范围均在 `allow` 中 → **自动放行**
+4. 否则 → 按 `defaultMode` 处理
+
+### 示例：宽松模式（默认）
+
+```json
+{
+  "permissions": {
+    "defaultMode": "allowAll"
+  }
+}
+```
+
+默认行为：所有操作自动放行，无需确认。
+
+### 示例：严格模式
+
+```json
+{
+  "permissions": {
+    "allow": ["read-in-cwd", "write-in-cwd", "query-git-log"],
+    "defaultMode": "askAll"
+  }
+}
+```
+
+此配置的效果：
+- 工作区内读写、Git 查询 → 自动放行
+- 其他操作都需要用户确认。
+
+
+## 持久化机制
+
+当用户在权限提示中选择 "Yes, and always allow" 后，对应的权限范围会被写入当前项目的 `.deepcode/settings.json` 文件中：
+
+- 新增范围会追加到 `permissions.allow` 列表
+- 如果该范围之前存在于 `deny` 或 `ask` 中，会被自动移除
+- 不会重复写入已存在的范围
+
+这样后续同类操作就不再询问。