zebbern/icron

zebbern のプロフィール

記録された最新のソース活動: 2026年2月6日 00:01

SkillsMP カタログ更新: 2026年10月17日 00:09

収集済み skills: 5

GitHub スター: 5
GitHub フォーク: 1

リポジトリの README

README の元ファイル：README.md

README の元ファイルを読む

README 全文を表示

<div align="center"> <img src="/case/icron.png" alt="https://github.com/zebbern/icron/blob/main/case/icron.png" width="400"> <kbd>icron YOUR lightweight Personal AI Assistant</kbd> </div> <a href="#install">Install</a> · <a href="#quick-start">Quick Start</a> · <a href="#key-features">Features</a> · <a href="#chat-channels">Channels</a> · <a href="#memory--persistence">Memory</a> · <a href="#reminders--scheduling">Reminders</a> <a href="#slash-commands">Commands</a> · <a href="#mcp-server-support">MCP</a> · <a href="#security">Security</a> · <a href="#configuration-reference">Config</a> · <a href="#cli-reference">CLI</a> · <a href="#docker">Docker</a> > [!note] > icron is a lightweight personal AI assistant built for simplicity and extensibility. > > - Multi-platform support (Discord, Telegram, WhatsApp, Slack, Feishu) > - Works with any OpenAI-compatible LLM (Anthropic, OpenRouter, Together, Groq, vLLM) > - Built-in tools: file operations, web search, shell execution, memory, scheduling > > Perfect for research, coding assistance, reminders, and daily automation. | Category | Tools | |----------|-------| | **File Operations** | `read_file`, `write_file`, `edit_file`, `list_dir`, `rename_file`, `move_file`, `copy_file`, `create_dir` | | **Code Search** | `glob` (find files), `grep` (search content) | | **Shell Execution** | `exec` (run commands with safety controls) | | **Web Access** | `web_search` (Brave API), `web_fetch` (extract content) | | **Memory** | `memory_search`, `memory_write`, `memory_get`, `memory_list` | | **Scheduling** | `reminder_set`, `reminder_list`, `reminder_cancel` | | **Slash Commands** | `/help`, `/sessions`, `/session`, `/remind`, `/search`, `/memory` | | **Screenshots** | `screenshot` (capture web pages with Playwright) | | **Subagents** | `spawn` (background task delegation) | | **MCP** | Connect external MCP servers for unlimited extensibility | ## Key Features - **Semantic Memory**: OpenClaw-style permanent Markdown memory with hybrid BM25 + vector search. Auto-detects embedding providers (OpenAI, Gemini, Ollama). - **Multi-Channel**: Chat through Discord, Telegram, WhatsApp, Slack, or Feishu with unified conversation history. - **LLM Flexibility**: Works with any OpenAI-compatible provider including local vLLM/Ollama. - **MCP Integration**: Extend capabilities by connecting external MCP servers. - **Skills System**: Modular task automation with built-in and custom skills. ## Install **From source** (recommended for development) ```bash git clone https://github.com/zebbern/icron.git cd icron pip install -e . ``` **From PyPI** (stable) ```bash pip install icron ``` ## Quick Start **1. Initialize** (choose one) ```bash # Guided setup wizard (recommended for first-time users) icron setup # Or quick initialization icron onboard ``` **2. Configure** (`~/.icron/config.json`) ```json { "providers": { "openrouter": { "apiKey": "sk-or-v1-xxx" } }, "agents": { "defaults": { "model": "anthropic/claude-sonnet-4-20250514" } } } ``` **3. Chat** ```bash icron agent -m "What is 2+2?" ``` ## Local Models (vLLM) Run icron with your own local models using vLLM or any OpenAI-compatible server. **1. Start your vLLM server** ```bash vllm serve meta-llama/Llama-3.1-8B-Instruct --port 8000 ``` **2. Configure** (`~/.icron/config.json`) ```json { "providers": { "vllm": { "apiKey": "dummy", "apiBase": "http://localhost:8000/v1" } }, "agents": { "defaults": { "model": "meta-llama/Llama-3.1-8B-Instruct" } } } ``` **3. Chat** ```bash icron agent -m "Hello from my local LLM!" ``` > The `apiKey` can be any non-empty string for local servers. ## Chat Channels Talk to icron through Telegram, WhatsApp, Discord, Slack, or Feishu. | Channel | Status | Setup | |---------|--------|-------| | **Discord** | ✅ Full support | Easy (bot token) | | **Telegram** | ✅ Full support + voice | Easy (bot token) | | **WhatsApp** | ✅ Basic support | Medium (QR scan) | | **Slack** | ✅ Full support | Medium (Socket Mode) | | **Feishu/Lark** | ✅ Full support | Medium (WebSocket) | <details> <summary>Discord Setup</summary> **1. Create a Discord bot** 1. Go to [Discord Developer Portal](https://discord.com/developers/applications) 2. Create new application, go to **Bot**, copy the token 3. Enable **Message Content Intent** 4. Go to **OAuth2 → URL Generator**, select `bot` scope with permissions: Send Messages, Read Messages, Read Message History 5. Invite bot to your server **2. Configure** ```json { "channels": { "discord": { "enabled": true, "token": "YOUR_BOT_TOKEN", "allowFrom": [YOUR_USER_ID], "allowedChannels": [] } } } ``` **3. Run** ```bash icron gateway ``` </details> <details> <summary>Telegram Setup</summary> **1. Create a bot** - Open Telegram, find `@BotFather` - Send `/newbot`, follow prompts - Copy the token **2. Configure** ```json { "channels": { "telegram": { "enabled": true, "token": "YOUR_BOT_TOKEN", "allowFrom": ["YOUR_USER_ID"] } } } ``` > Get your user ID from `@userinfobot` on Telegram. **3. Run** ```bash icron gateway ``` > **Voice transcription**: If you configure a Groq API key, voice messages will be automatically transcribed via Whisper. </details> <details> <summary>WhatsApp Setup</summary> Requires **Node.js ≥18**. **1. Start the bridge** ```bash icron channels login # Scan QR with WhatsApp → Settings → Linked Devices ``` **2. Configure** ```json { "channels": { "whatsapp": { "enabled": true, "allowFrom": ["+1234567890"] } } } ``` **3. Run** (two terminals) ```bash # Terminal 1 - WhatsApp bridge icron channels login # Terminal 2 - Gateway icron gateway ``` </details> ## Web UI The gateway provides a web interface for configuration. ```bash icron gateway # Open http://localhost:3883/app ``` **Quick Settings UI:** - Model selection - API keys (Together, OpenRouter, Brave Search) - Channel toggles (Telegram, WhatsApp, Discord) - Security settings (restrict tools to workspace) - Context limits (max tokens for conversation history) ## Memory & Persistence icron has persistent memory that survives restarts: ``` # Store a fact "Remember that my project deadline is January 31st" # Recall later "When is my project deadline?" # List all memories "What do you remember about me?" ``` Memory is stored in `~/.icron/workspace/memory/` as markdown files. ## Reminders & Scheduling **Set reminders naturally:** ``` "Remind me in 30 minutes to take a break" "Remind me at 2pm to call John" "Remind me tomorrow at 9am about the meeting" ``` **Or use slash commands for quick reminders:** ``` /remind 30m take a break /remind 2h call john ``` **Cron jobs for recurring tasks:** ```bash # Add a daily reminder icron cron add --name "morning" --message "Good morning!" --cron "0 9 * * *" # Add a one-time reminder icron cron add --name "meeting" --message "Meeting starts!" --at "2025-01-31T15:00:00" # List and manage jobs icron cron list icron cron remove <job_id> ``` ## Slash Commands Quick commands that bypass the LLM for instant response: | Command | Description | |---------|-------------| | `/help` | List all commands | | `/help [topic]` | Detailed help (sessions, memory, reminders, search) | | `/sessions` | List all sessions | | `/session clear` | Clear current session | | `/session new` | Start fresh session | | `/session rename [name]` | Rename session | | `/session switch [id]` | Switch sessions | | `/remind [time] [message]` | Set quick reminder | | `/search [query]` | Quick web search | | `/memory` | Access memory | | `/skills` | List available skills | | `/skills run [name]` | Run a specific skill | | `/weather [location]` | Get weather for a location | | `/templates` | List available templates | | `/template [name]` | Run a template by name | | `/collab [task]` | Multi-model collaboration | ## MCP Server Support icron integrates with [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) servers for unlimited extensibility. Connect to file systems, databases, APIs, and more. ### Quick Setup ```json { "tools": { "mcp": { "enabled": true, "servers": { "time": { "transport": "stdio", "command": "npx", "args": ["-y", "@guanxiong/mcp-server-time"] }, "filesystem": { "transport": "stdio", "command": "npx", "args": ["-y", "@anthropic/mcp-server-filesystem", "/"] } } } } } ``` ### Popular MCP Servers | Server | Package | Tools | API Key | |--------|---------|-------|---------| | **Time** | `@guanxiong/mcp-server-time` | Get time, convert timezones | No | | **Filesystem** | `@anthropic/mcp-server-filesystem` | Read/write/search files | No | | **Fetch** | `mcp-fetch-server` | HTTP requests, web scraping | No | | **Sequential Thinking** | `@modelcontextprotocol/server-sequential-thinking` | Step-by-step reasoning | No | | **GitHub** | `@modelcontextprotocol/server-github` | Issues, PRs, repos | `GITHUB_PERSONAL_ACCESS_TOKEN` | | **Brave Search** | `@anthropic/mcp-server-brave-search` | Web search | `BRAVE_API_KEY` | | **PostgreSQL** | `@anthropic/mcp-server-postgres` | Database queries | Connection string | | **Slack** | `@anthropic/mcp-server-slack` | Slack messages | `SLACK_BOT_TOKEN` | ### Server Types **Standard (stdio) - npm packages:** ```json { "time": { "transport": "stdio", "command": "npx", "args": ["-y", "@guanxiong/mcp-server-time"] } } ``` **Python servers:** ```json { "custom": { "transport": "stdio", "command": "python", "args": ["/path/to/mcp_server.py"] } } ``` **SSE (Server-Sent Events) - remote servers:** ```json { "remote": { "transport": "sse", "url": "https://mcp.example.com/sse" } } ``` ### Servers with API Keys Set environment variables before running icron: ```bash export GITHUB_PERSONAL_ACCESS_TOKEN="ghp_xxx" export BRAVE_API_KEY="BSA_xxx" icron gateway ``` Or in Docker: ```bash docker run -e GITHUB_PERSONAL_ACCESS_TOKEN="ghp_xxx" ... ``` ### Web UI Configuration The **Settings → MCP** tab provides 19 pre-configured MCP server presets. Enable servers with one click and they'll be added to your config automatically. MCP tools appear alongside built-in tools and are available to the agent immediately after configuration. ## Security **Workspace restriction:** ```json { "tools": { "exec": { "restrictToWorkspace": true } } } ``` When enabled, file operations are limited to `~/.icron/workspace/`. **Built-in safety:** - Dangerous shell commands are blocked (rm -rf, format, dd, etc.) - Command execution has configurable timeout (default 60s) - Output is truncated to prevent context overflow ## Configuration Reference Config file: `~/.icron/config.json` | Section | Key | Description | |---------|-----|-------------| | `agents.defaults.model` | string | Default LLM model | | `providers.openrouter.apiKey` | string | OpenRouter API key | | `providers.anthropic.apiKey` | string | Anthropic API key | | `providers.together.apiKey` | string | Together AI API key | | `providers.groq.apiKey` | string | Groq API key (for voice) | | `providers.openai.apiKey` | string | OpenAI API key | | `providers.gemini.apiKey` | string | Google Gemini API key | | `providers.zhipu.apiKey` | string | Zhipu AI API key | | `tools.exec.timeout` | int | Shell command timeout (seconds) | | `tools.exec.restrictToWorkspace` | bool | Limit file access to workspace | | `tools.exec.maxContextTokens` | int | Max tokens for conversation history | | `tools.web.search.apiKey` | string | Brave Search API key | | `tools.mcp.enabled` | bool | Enable MCP servers | | `channels.discord.enabled` | bool | Enable Discord | | `channels.telegram.enabled` | bool | Enable Telegram | | `channels.whatsapp.enabled` | bool | Enable WhatsApp | <details> <summary>Full config example</summary> ```json { "agents": { "defaults": { "model": "anthropic/claude-sonnet-4-20250514", "maxToolIterations": 20 } }, "providers": { "openrouter": { "apiKey": "sk-or-v1-xxx" }, "groq": { "apiKey": "gsk_xxx" } }, "channels": { "discord": { "enabled": true, "token": "xxx", "allowFrom": [123456789], "allowedChannels": [] }, "telegram": { "enabled": false, "token": "123456:ABC...", "allowFrom": ["123456789"] }, "whatsapp": { "enabled": false, "allowFrom": ["+1234567890"] } }, "tools": { "exec": { "timeout": 60, "restrictToWorkspace": false, "maxContextTokens": 100000 }, "web": { "search": { "apiKey": "BSA..." } }, "mcp": { "enabled": false, "servers": {} } } } ``` </details> ## CLI Reference | Command | Description | |---------|-------------| | `icron setup` | Guided setup wizard (recommended) | | `icron onboard` | Initialize config & workspace | | `icron validate` | Validate configuration file | | `icron agent -m "..."` | Send a message | | `icron agent` | Interactive chat mode | | `icron gateway` | Start gateway (web UI + channels) | | `icron status` | Show status | | `icron channels login` | Link WhatsApp (QR scan) | | `icron channels status` | Show channel status | | `icron cron list` | List scheduled jobs | | `icron cron add` | Add a scheduled job | | `icron cron remove <id>` | Remove a scheduled job | ## Docker ```bash # Build the image docker build -t icron . # Initialize config docker run -v ~/.icron:/root/.icron --rm icron onboard # Edit config vim ~/.icron/config.json # Run gateway docker run -v ~/.icron:/root/.icron -p 3883:3883 icron gateway ``` **Railway deployment:** - Uses `Dockerfile` and `railway.json` - Set environment variables: `TOGETHER_API_KEY`, `MODEL`, `ICRON_WRITE_CONFIG=1` - Optional: `TELEGRAM_TOKEN`, `TELEGRAM_ALLOW_FROM`, `WEBSEARCH_API_KEY` ## Project Structure ``` icron/ ├── agent/ # Core agent logic │ ├── loop.py # Agent loop (LLM ↔ tools) │ ├── context.py # Prompt builder │ ├── memory.py # Persistent memory │ ├── subagent.py # Background task execution │ └── tools/ # Built-in tools ├── channels/ # Discord, Telegram, WhatsApp, Slack, Feishu ├── bus/ # Message routing ├── cron/ # Scheduled tasks ├── heartbeat/ # Proactive checks ├── providers/ # LLM providers ├── session/ # Conversation sessions ├── mcp/ # MCP server integration ├── config/ # Configuration └── cli/ # CLI commands ``` ## Documentation (DRY) Keep documentation in `/docs` for reusability: | File | Purpose | |------|---------| | `architecture.md` | System overview, components, data flow | | `decisions.md` | Why things were built this way | | `conventions.md` | Naming, formatting, repo rules | | `workflows.md` | How things are built, tested, deployed | | `integrations.md` | APIs, services, auth, rate limits | | `glossary.md` | Domain terms and meanings | > **Rule of thumb:** If you explain it twice, document it once. ## Contribute PRs welcome! The codebase is intentionally small and readable. **Roadmap:** - [x] Discord integration - [x] Voice transcription (Telegram/Groq) - [x] Persistent memory - [x] MCP server support - [x] Scheduled reminders - [x] Context trimming - [ ] WhatsApp voice transcription - [ ] Multi-modal (images, voice) - [ ] Email integration - [ ] Calendar integration icron is for educational and personal use

リポジトリの README を読み込んでいます…

このリポジトリの skills

ソフトウェア開発者ネットワーク・コンピュータシステム管理者市場調査アナリスト・マーケティングスペシャリスト

3 件の職業カテゴリ · 100% 分類済み

skill

職業分類

説明

更新

cron

ネットワーク・コンピュータシステム管理者

Schedule reminders and recurring tasks.

原文の言語: 英語

2026-02-06

github

ソフトウェア開発者

Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.

原文の言語: 英語

2026-02-06

summarize

市場調査アナリスト・マーケティングスペシャリスト

Summarize or extract text/transcripts from URLs, podcasts, and local files (great fallback for “transcribe this YouTube/video”).

原文の言語: 英語

2026-02-06

tmux

ネットワーク・コンピュータシステム管理者

Remote-control tmux sessions for interactive CLIs by sending keystrokes and scraping pane output.

原文の言語: 英語

2026-02-06

weather

ソフトウェア開発者

Get current weather and forecasts (no API key required).

原文の言語: 英語

2026-02-06

収集済み skill 5 件中 5 件を表示しています。

AI アシスタントでインストール

このプロンプトをコピーして、使用中の AI アシスタントに送ってください。

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/zebbern/icron.