orquesta-cli 0.1.12

package/.eslintrc.json

@@ -0,0 +1,26 @@
+{
+  "parser": "@typescript-eslint/parser",
+  "parserOptions": {
+    "ecmaVersion": 2022,
+    "sourceType": "module",
+    "project": "./tsconfig.json"
+  },
+  "plugins": ["@typescript-eslint"],
+  "extends": [
+    "eslint:recommended",
+    "plugin:@typescript-eslint/recommended",
+    "plugin:@typescript-eslint/recommended-requiring-type-checking"
+  ],
+  "rules": {
+    "@typescript-eslint/explicit-function-return-type": "warn",
+    "@typescript-eslint/no-unused-vars": ["error", { "argsIgnorePattern": "^_" }],
+    "@typescript-eslint/no-explicit-any": "warn",
+    "@typescript-eslint/no-floating-promises": "error",
+    "no-console": "off"
+  },
+  "env": {
+    "node": true,
+    "es2022": true
+  },
+  "ignorePatterns": ["dist/", "node_modules/", "*.js"]
+}

package/.prettierrc.json

@@ -0,0 +1,10 @@
+{
+  "semi": true,
+  "trailingComma": "es5",
+  "singleQuote": true,
+  "printWidth": 100,
+  "tabWidth": 2,
+  "useTabs": false,
+  "arrowParens": "always",
+  "endOfLine": "lf"
+}

package/CLAUDE.md

@@ -0,0 +1,199 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+Local-CLI is an OpenAI-compatible local CLI coding agent for offline enterprise environments. It works with any OpenAI-compatible API (vLLM, Ollama, LM Studio, Azure OpenAI, Google Gemini).
+
+## Build, Test, and Lint Commands
+
+```bash
+# Build
+npm run build          # Compile TypeScript
+npm run watch          # Watch mode compilation
+npm run dev            # Run with ts-node (development)
+npm run start          # Run compiled dist/cli.js
+
+# Test
+npm run test           # Full pytest suite (tests/test_eval.py)
+npm run test:quick     # Quick tests (skip slow tests)
+
+# Code Quality
+npm run lint           # Run ESLint on src/**/*.ts
+npm run lint:fix       # Auto-fix ESLint issues
+npm run format         # Format with Prettier
+npm run prepr          # Pre-PR check: lint + build
+
+# Electron (optional desktop app)
+npm run electron:dev   # Electron development mode
+npm run electron:build # Build Electron app
+npm run electron:package # Package for Windows (creates app.asar)
+```
+
+## Electron Build & Deploy (IMPORTANT)
+
+**빌드 후 Windows 배포 시 반드시 확인:**
+
+1. `npm run electron:build` - renderer/main/preload 빌드 (dist-electron/)
+2. `npm run electron:package` - app.asar 생성 (release/win-unpacked/resources/)
+3. `cp release/win-unpacked/resources/app.asar /mnt/c/LOCAL-CLI/resources/` - Windows에 배포
+
+**주의사항:**
+- `electron:build`만 실행하면 app.asar가 업데이트되지 않음
+- Windows의 app.asar 파일 시간을 확인해서 최신 빌드인지 반드시 검증
+- CSS 변경이 반영 안되면 app.asar 배포 여부 먼저 확인
+- **IMPORTANT**: `npm run electron:package`는 WSL에서 wine 오류가 발생하지만, app.asar는 오류 전에 이미 생성됨. wine 오류는 서명 단계에서 발생하므로 무시해도 됨
+- **절대 금지**: `npx asar pack`으로 수동 패키징하면 node_modules 의존성이 빠져서 앱이 실행 안됨. 반드시 electron-builder가 생성한 app.asar 사용
+
+## WSL에서 Windows Electron 앱 실행하기 (IMPORTANT)
+
+**잘못된 방법 (절대 사용 금지):**
+```bash
+# 이 방법들은 모두 실패함:
+cmd.exe /c "start C:\LOCAL-CLI\..."      # UNC 경로 오류
+cmd.exe /c 'C:\LOCAL-CLI\...'            # 작업 디렉토리가 WSL UNC 경로라서 실패
+"/mnt/c/LOCAL-CLI/앱이름.exe" &          # 백그라운드에서 조용히 실패
+```
+
+**왜 실패하는가:**
+- WSL에서 cmd.exe를 실행하면 현재 작업 디렉토리가 `\\wsl.localhost\Ubuntu-22.04\...` (UNC 경로)로 설정됨
+- CMD.EXE는 UNC 경로를 작업 디렉토리로 지원하지 않음
+- 경로에 공백이 있으면 이스케이핑이 복잡해져서 추가 오류 발생
+
+**올바른 방법:**
+```bash
+# PowerShell 사용 (권장)
+powershell.exe -Command "& 'C:\LOCAL-CLI\LOCAL-CLI PowerShell UI.exe'"
+
+# 또는 Start-Process 사용
+powershell.exe -Command "Start-Process 'C:\LOCAL-CLI\LOCAL-CLI PowerShell UI.exe'"
+```
+
+**앱 실행 확인:**
+```bash
+# 프로세스 확인
+tasklist.exe | grep -i "LOCAL-CLI\|Electron"
+```
+
+**주의:** 앱이 실행 후 바로 종료되면 앱 자체의 초기화 오류일 가능성이 높음. 로그 확인 필요.
+
+## Architecture
+
+### Plan & Execute Pattern
+
+The core execution model automatically breaks user requests into TODO lists and executes them sequentially:
+
+```
+User Input → Planning LLM (generates TODOs) → Plan Executor → Per-task Agent → Tool Execution → Result
+```
+
+Key files:
+- `src/orchestration/plan-executor.ts` - Core execution logic (React-independent pure logic)
+- `src/agents/planner/` - TODO list generation from user requests
+- `src/ui/components/PlanExecuteApp.tsx` - Main interactive UI component
+
+### Tool System (6 Categories)
+
+Tools are registered in `src/tools/registry.ts` with multi-category support:
+
+| Category | Description |
+|----------|-------------|
+| LLM Simple | Called via tool_calls, no sub-LLM (file ops, bash) |
+| LLM Agent | Called via tool_calls, uses sub-LLM (docs search) |
+| System Simple | Auto-triggered by logic, no sub-LLM |
+| System Agent | Auto-triggered by logic, uses sub-LLM |
+| User Commands | Slash commands (/help, /settings, etc.) |
+| MCP Tools | Model Context Protocol tools |
+
+### Directory Structure
+
+```
+src/
+├── cli.ts                    # CLI entry point (Commander.js)
+├── constants.ts              # Global paths (~/.local-cli/)
+├── core/
+│   ├── llm/llm-client.ts    # OpenAI-compatible API wrapper (Axios)
+│   ├── config/              # Settings file management
+│   ├── session/             # Session persistence
+│   └── compact/             # Auto-context compression at 80% capacity
+├── agents/
+│   ├── planner/             # TODO list generation
+│   └── docs-search/         # LLM-based documentation search
+├── orchestration/
+│   └── plan-executor.ts     # Main Plan & Execute logic
+├── prompts/                  # Centralized prompt management
+│   ├── shared/              # Language/tool/codebase rules
+│   └── agents/              # Agent-specific prompts
+├── tools/
+│   ├── llm/simple/          # File, bash, user interaction tools
+│   ├── browser/             # Chrome/Edge automation (optional)
+│   └── office/              # Word, Excel, PowerPoint (optional, Windows)
+├── ui/
+│   ├── components/          # Ink React components
+│   └── hooks/               # Custom React hooks
+├── errors/                   # Typed error classes
+└── utils/
+    ├── logger.ts            # JSON stream logging
+    └── platform-utils.ts    # OS detection (Windows, WSL, macOS, Linux)
+```
+
+### LLM Client
+
+`src/core/llm/llm-client.ts` handles all API communication:
+- OpenAI-compatible endpoint support
+- Streaming and non-streaming responses
+- Auto-retry logic (up to 3 attempts)
+- Model-specific preprocessing for reasoning LLMs
+- 10-minute timeout for long requests
+
+### Configuration
+
+- Home directory: `~/.local-cli/`
+- Config file: `~/.local-cli/config.json`
+- Managed by: `src/core/config/config-manager.ts`
+
+## Adding New Tools
+
+1. Create tool in appropriate directory under `src/tools/`
+2. Export tool definition with `name`, `description`, `parameters`, and `execute` function
+3. Register in `src/tools/registry.ts` with appropriate categories
+4. Tools requiring approval: set `requiresApproval: true`
+
+Tool definition pattern:
+```typescript
+export default [{
+  name: 'tool_name',
+  description: 'What the tool does',
+  parameters: { /* JSON Schema */ },
+  execute: async (params, context) => { /* implementation */ }
+}];
+```
+
+## Adding Slash Commands
+
+Slash commands are handled in `src/core/slash-command-handler.ts`. Register new commands there following the existing pattern.
+
+## Debugging
+
+CLI flags:
+- `--verbose` - Verbose output
+- `--debug` - Debug mode
+- `--llm-log` - Log LLM requests/responses
+- `--eval` - Evaluation mode (NDJSON event streaming)
+
+JSON stream logs are written by `src/utils/logger.ts` for analysis.
+
+## Platform Notes
+
+- **WSL2**: Browser and Office tools work via network mirroring to Windows host
+- **Windows**: Office tools use COM automation (requires installed Office)
+- **Browser tools**: Require Chrome or Edge installed
+
+## Key Patterns
+
+- Single tool execution per LLM invocation (enforced)
+- Prompts centralized in `src/prompts/` (no scattered prompts)
+- Plan executor is React-independent for testability
+- TypeScript strict mode enabled with ESM modules
+- Path alias: `@/*` maps to `src/*`

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 HanSyngha
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,280 @@
+# orquesta-cli
+
+**AI-powered coding assistant with local LLM support** - the self-hosted alternative to Claude CLI.
+
+> Use your own LLM (vLLM, Ollama, LM Studio, Azure OpenAI, or any OpenAI-compatible API) as a full coding agent - no cloud dependency, no subscription costs.
+
+## ⚠️ Security Warning
+
+**IMPORTANT:** Orquesta CLI gives AI models the ability to execute commands on your system. Please read [SECURITY.md](./SECURITY.md) before use.
+
+**Quick Security Tips:**
+- ✅ Use only trusted LLM endpoints
+- ✅ Enable manual approval for production (`autoApprove: false`)
+- ✅ Run in isolated environments (Docker/VM)
+- ✅ Review commands before execution
+- ❌ Never expose sensitive credentials as environment variables
+
+---
+
+## Why orquesta-cli?
+
+**Freedom from cloud dependencies** - Run your own LLM infrastructure:
+
+| Feature | Claude CLI | orquesta-cli |
+|---------|-----------|--------------|
+| **LLM Provider** | Anthropic (cloud) | Any OpenAI-compatible endpoint |
+| **Hosting** | Subscription required | Self-hosted / on-premises |
+| **Options** | Claude only | vLLM, Ollama, LM Studio, Azure OpenAI, local models |
+| **Cost** | Per-token pricing | Your infrastructure |
+| **Privacy** | Data sent to Anthropic | Data stays local |
+| **Integration** | - | Works with orquesta-agent |
+
+---
+
+## Key Features
+
+| Benefit | Description |
+|---------|-------------|
+| **Zero Cloud Dependency** | Runs entirely on your local/on-prem LLM. Your code never leaves your network. |
+| **No API Cost** | Use open-source models (Llama, Qwen, DeepSeek, etc.) for free. |
+| **Any OpenAI-Compatible API** | Works with vLLM, Ollama, LM Studio, Azure OpenAI, Google Gemini, and more. |
+| **Autonomous Coding Agent** | Reads, searches, edits, and creates code files - not just chat. |
+| **Plan & Execute** | Breaks complex tasks into TODO steps and executes them sequentially. |
+| **Safe by Default** | Supervised mode requires your approval before any file modification. |
+| **Office Automation** | Control Excel, Word, PowerPoint directly via PowerShell/COM (Windows). |
+| **Browser Automation** | Chrome/Edge CDP control - navigate, click, screenshot, scrape data. |
+| **orquesta-agent Compatible** | Works seamlessly with orquesta-agent for remote execution |
+
+---
+
+## Installation
+
+```bash
+npm install -g orquesta-cli
+```
+
+Or run directly with npx:
+
+```bash
+npx orquesta-cli
+```
+
+---
+
+## Quick Start
+
+### Interactive Mode
+
+```bash
+orquesta
+```
+
+This launches the full interactive UI with plan & execute, file browser, session history, etc.
+
+### One-Shot Prompt Execution
+
+```bash
+orquesta -p "fix the bug in auth.ts"
+```
+
+This executes the prompt and exits - compatible with `orquesta-agent` for remote execution.
+
+---
+
+## Configuration
+
+```bash
+orquesta            # Setup wizard on first run
+/settings           # Settings menu while running
+```
+
+Any OpenAI-compatible API works:
+- **vLLM**: `http://localhost:8000/v1`
+- **Ollama**: `http://localhost:11434/v1`
+- **LM Studio**: `http://localhost:1234/v1`
+- **Azure OpenAI**: `https://YOUR_RESOURCE.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT`
+- **Google Gemini**: Compatible with OpenAI SDK
+- **Any internal LLM server**
+
+Configuration is stored in `~/.orquesta-cli/config.json`.
+
+---
+
+## Usage with orquesta-agent
+
+orquesta-cli works seamlessly with [orquesta-agent](https://www.npmjs.com/package/orquesta-agent) for remote execution:
+
+```bash
+# On your server/VM
+npm install -g orquesta-cli
+
+# Configure your local LLM endpoint
+orquesta  # Follow setup wizard
+
+# Run orquesta-agent (it will use orquesta-cli instead of claude)
+npx orquesta-agent --token <your-token>
+```
+
+orquesta-agent will execute prompts using `orquesta-cli` with your local LLM.
+
+---
+
+## Commands & Features
+
+### Plan & Execute
+Automatically decomposes requests into TODO steps and executes them:
+
+```
+You: Add a logging system to the project
+
+TODO List                            1/3
+  [x] Create logger.ts
+  [ ] Add imports to existing code
+  [ ] Apply error handling
+```
+
+### Supervised Mode
+Every file modification requires your explicit approval:
+
+- **Tab** to toggle between Auto / Supervised mode
+- Only file modification tools need approval (read/search are always allowed)
+- Reject with feedback to guide the agent's next attempt
+
+### Slash Commands
+| Command | Description |
+|---------|-------------|
+| `/help` | Help |
+| `/clear` | Reset conversation |
+| `/compact` | Compress conversation |
+| `/load` | Load saved session |
+| `/model` | Switch model |
+| `/settings` | Settings menu |
+| `/usage` | Token usage |
+| `/docs` | Docs management |
+| `/tool` | Toggle optional tools (browser/office) |
+
+### Keyboard Shortcuts
+| Key | Action |
+|-----|--------|
+| `Ctrl+C` | Exit |
+| `ESC` | Interrupt current task |
+| `Tab` | Toggle Auto / Supervised |
+| `@` | File browser |
+| `/` | Command autocomplete |
+
+---
+
+## Command-Line Options
+
+```bash
+orquesta [options]
+
+Options:
+  -p, --print <prompt>           Execute a prompt and exit (non-interactive)
+  --dangerously-skip-permissions Skip all permission prompts (auto-approve)
+  --verbose                      Enable verbose logging
+  --debug                        Enable debug logging
+  --llm-log                      Log LLM requests/responses
+  --eval                         Evaluation mode (NDJSON event streaming)
+  -V, --version                  Output version number
+```
+
+---
+
+## Office Automation (60+ Tools)
+
+**Windows only** - Control Microsoft Office via PowerShell/COM:
+
+| App | Capabilities |
+|-----|-------------|
+| **Excel** | Read/write cells, create charts, formatting, conditional formatting, pivot tables, formulas |
+| **Word** | Write text, headers, tables, images, footnotes, find/replace, styles, TOC |
+| **PowerPoint** | Create slides, add text/images/shapes, apply themes, speaker notes |
+
+---
+
+## Browser Automation
+
+Chrome/Edge automation via Chrome DevTools Protocol:
+- Navigate pages, click elements, fill forms
+- Take screenshots, extract text
+- No external server required
+
+---
+
+## Session Management
+
+- Save and restore conversation history
+- Auto-context compression at 80% capacity
+- Resume work exactly where you left off
+
+---
+
+## Development
+
+```bash
+# Clone the repository
+git clone https://github.com/orquesta/orquesta.git
+cd orquesta/packages/orquesta-cli
+
+# Install dependencies
+npm install
+
+# Build
+npm run build
+
+# Run locally
+npm start
+# Or: node dist/cli.js
+```
+
+---
+
+## Requirements
+
+- Node.js v20+
+- npm v10+
+- Git
+- Windows (for Office/Browser automation via PowerShell)
+
+---
+
+## Comparison with Claude CLI
+
+Both tools are **AI-powered coding assistants** but differ in architecture:
+
+### Claude CLI (Official)
+- ✅ Direct Anthropic API access
+- ✅ Claude 4.5/4.6 models
+- ✅ Official support
+- ❌ Requires subscription
+- ❌ Cloud-only
+- ❌ Data sent to Anthropic
+
+### orquesta-cli
+- ✅ Self-hosted LLM support
+- ✅ Any OpenAI-compatible model
+- ✅ Full data privacy
+- ✅ No subscription needed (BYO compute)
+- ✅ Plan & Execute architecture
+- ✅ Office/Browser automation
+- ❌ Model quality depends on your LLM choice
+
+**Use Claude CLI if:** You want the best model quality and have a subscription.
+
+**Use orquesta-cli if:** You need data privacy, self-hosting, or want to use specific models (local, on-prem, etc.).
+
+---
+
+## License
+
+MIT
+
+---
+
+## Links
+
+- **Dashboard**: https://orquesta.live
+- **GitHub**: https://github.com/orquesta/orquesta
+- **npm**: https://www.npmjs.com/package/orquesta-cli