Skip to content

RedPlanetHQ/core-cli

Repository files navigation

Core CLI - AI-Powered Task Management That Never Forgets

An intelligent task management system that remembers your entire world - your meetings, your tasks, your team discussions, your commitments - and seamlessly integrates with all your tools.

Unlike traditional task managers or coding-focused CLIs like claude-code, Core CLI is built for task management with a persistent knowledge graph that connects your calendar, GitHub issues, Linear tickets, Slack messages, and conversations into one intelligent context.

The Problem: Tasks scattered across GitHub, Linear, Slack, and your notes. No single view. No context. No memory.

Core CLI: One unified task hub. All integrations. Forever remembered.

npm version License: AGPL v3

What Makes Core Different

Core CLI is an AI-powered task management system with a beautiful GUI interface (similar to claude-code), built specifically for developers who want intelligent task organization without leaving the terminal.

What Can You Do?

Task Management

# View your current tasks
/tasks

# Quick task entry mode
--
# β†’ Enter task mode for rapid task creation

# Natural language task updates
"Mark the Linear sync task as complete"
"Move the GitHub PR review to next week"
"Update the API design task to high priority"

Integration-Powered Task Sync

# Smart routine to sync across all your tools
@sync
# β†’ Checks GitHub, Linear, Slack for updates
# β†’ Creates tasks for items needing attention
# β†’ Updates existing tasks with changes
# β†’ Helps you prioritize what matters

# Task queries with context
"What should I focus on this week?"
# β†’ Analyzes your calendar, sprint commitments, GitHub PRs, Linear tickets

# Meeting prep with task context
"Prep me for standup"
# β†’ Summarizes tasks completed, blockers, references team discussions

Document-Aware Task Management

# Reference documents created in Core
"@linear-product-doc write the following features as tasks"
"@meeting-notes create tasks from action items"
"@sprint-planning-doc add these to my weekly tasks"

# Context-aware queries
"What did we decide about the API redesign last week?"
"Find all tasks related to the authentication feature"
"What commitments did I make in Monday's meeting?"

Intelligent Task Organization

# Automatic weekly organization
# Tasks stored in weekly files: 2025-W48.md
# Incomplete tasks automatically roll over to next week

# Time-aware prioritization
"I have 3 hours before my 2pm meeting, what tasks should I tackle?"
# β†’ Considers your calendar, task priorities, and past completion patterns

How It Works

Your World                Knowledge Graph              Your AI Tools
───────────              ────────────────────              ─────────────
Gmail          ─────▢                            ─────▢    CLI
Calendar       ─────▢    Temporal Memory         ─────▢    IDE Extensions
GitHub         ─────▢    88% Recall Accuracy     ─────▢    Browser
Slack          ─────▢    Cross-Domain Links      ─────▢    Mobile
Linear         ─────▢    Entity Relationships    ─────▢    Any LLM
Google Docs    ─────▢    Event Timelines         ─────▢    API
Sheets         ─────▢    Smart Clustering        ─────▢    Everywhere

One memory layer. All your context. Every AI tool.

Key Features

  • AI-Powered Task Management - Update and manage tasks using natural language
  • Persistent Memory - Temporal knowledge graph that remembers all context and conversations
  • Weekly Task Files - Automatic markdown files (2025-W48 format) organized by week
  • Smart Task Rollover - Incomplete tasks automatically move to next week
  • Quick Task Mode - Use -- to instantly enter rapid task creation mode
  • Integration Hub - Native GitHub, Linear, and Slack integrations for unified task management
  • Smart Sync Routines - Built-in routines check integrations, create tasks, and help prioritize
  • Document References - Use @ command to reference and work with Core documents
  • Conversation Linking - All tasks automatically linked to conversations for full context
  • CLI-First Design - Fast, efficient task management directly from your terminal
  • Data Ownership - You own your task history and knowledge graph

Installation

Quick Start

# Install globally
pnpm install -g @redplanethq/core-cli

# Start the CLI
core-cli

Or use directly with pnpm:

pnpm dlx @redplanethq/core-cli

First Run Setup

On first run, the configuration wizard will guide you through:

  1. Choose your AI provider - OpenAI, Anthropic, or other supported models
  2. Configure API keys - Set up your preferred LLM
  3. Connect integrations - Link Gmail, Calendar, GitHub, Linear, etc.
  4. Build your knowledge graph - Core ingests your past conversations and data
  5. Customize preferences - Set your assistant name, theme, and defaults

Configuration files are stored in your system's standard config directory.

Environment Variables

Set your AI provider API keys as environment variables:

# For OpenAI
export OPENAI_API_KEY=your_openai_key_here

# For Anthropic
export ANTHROPIC_API_KEY=your_anthropic_key_here

# For Core Memory (if using hosted Core)
export CORE_API_KEY=your_core_api_key_here

Requirements

  • Node.js >= 20
  • Active internet connection for AI provider access
  • (Optional) Core account for hosted knowledge graph

Features

Interactive Chat Interface

Conversational AI assistant directly in your terminal with real-time streaming responses.

Multiple AI Providers

Support for various AI models:

  • OpenAI (GPT-4, GPT-4 Turbo, GPT-3.5)
  • Anthropic (Claude 3.5 Sonnet, Claude 3 Opus)
  • More providers coming soon

MCP (Model Context Protocol) Integration

Enhanced capabilities through MCP servers:

  • Integrate with external systems and APIs
  • Extend functionality with custom MCP servers
  • Connect to GitHub, Linear, Slack, and more

Token Usage Tracking

Monitor your token consumption across conversations and sessions.

Commands

Chat Commands

Command Description
/help Display help information and available commands
/tasks View your current tasks
/sessions View all active coding agent sessions
-- Enter quick task mode for rapid task creation
sync Run smart routine to sync integrations and prioritize
/clear Clear conversation history (memory persists)
/model Select a different AI model
/provider Change AI provider
/mcp Manage MCP (Model Context Protocol) servers
/usage View token usage statistics
/set-name Customize your assistant's name

CLI Commands

Task Management

# List all tasks
core-cli tasks list

# Add a new task
core-cli tasks add "Implement user authentication API"

# Update an existing task
core-cli tasks update 42 "Updated task description"

# Delete a task
core-cli tasks delete 42

# View help
core-cli tasks help

Session Management

# List all active coding agent sessions
core-cli sessions list

# Attach to a session
core-cli sessions attach task-42-abcde
core-cli sessions attach 42  # If only one session for task 42

# Delete a specific session
core-cli sessions delete task-42-abcde

# Clear all sessions
core-cli sessions clear

# View help
core-cli sessions help

Coding Agents & Task Delegation

Core CLI can delegate implementation work to AI coding agents like Claude Code, Cursor, Aider, and others. These agents run in isolated tmux sessions with their own git worktrees, allowing you to work on multiple tasks in parallel without conflicts.

Supported Coding Agents

Core CLI supports the following coding agents out of the box:

Agent Description Installation
Claude Code Anthropic's official CLI for Claude npm install -g @anthropic/claude-code
Cursor AI AI-first code editor Download from cursor.sh
Aider AI pair programming in your terminal pip install aider-chat
Codex CLI OpenAI Codex command-line interface Custom installation

Setting Default Coding Agent

You can set your preferred coding agent in the CLI:

# Inside core-cli chat
/set-coding-agent

# Or ask naturally
"Set my default coding agent to Claude Code"

Task Delegation Workflow

1. Using Natural Language (Recommended)

The AI assistant can automatically launch coding agents for you:

# Inside core-cli chat
"Implement task 42"
"Start working on task 15 using Cursor"
"Delegate the API implementation to Aider"

The assistant will:

  • Validate the task exists
  • Launch the specified coding agent in a tmux session
  • Create an isolated git worktree for the task
  • Pass task context to the agent
  • Show you how to attach to the session

2. Using the @implement Routine

Core CLI has a built-in routine for task delegation:

# Inside core-cli chat
@implement 42           # Implement task #42 with default agent
@implement 15 cursor    # Implement task #15 with Cursor

3. Programmatic Access (Advanced)

The AI assistant has access to coding session tools and can:

  • Launch sessions with custom context
  • List active sessions
  • Close or delete sessions
  • Monitor session status

How Git Worktrees Work

When you delegate a task, Core CLI automatically:

  1. Creates a new git branch named task-{number}-{random} (e.g., task-42-abcde)
  2. Creates a git worktree in ../worktrees/task-42-abcde/
  3. Launches the coding agent in that worktree directory
  4. Runs in isolated tmux session so it doesn't interfere with your main work

Benefits:

  • Work on multiple tasks in parallel without git conflicts
  • Each task has its own clean workspace
  • Easy to switch between tasks
  • Automatic cleanup when session is deleted

Managing Active Sessions

List Sessions

# Inside core-cli chat
/sessions

# Or via CLI
core-cli sessions list

Output shows:

  • Agent name (e.g., [Claude Code])
  • Session name (e.g., task-42-abcde)
  • Task description preview
  • Session status and start time
  • Git worktree location
  • Attach command

Attach to a Session

# Via CLI
core-cli sessions attach task-42-abcde
core-cli sessions attach 42  # If only one session for this task

# Inside tmux session
# Press Ctrl+B then D to detach

Delete a Session

Deleting a session will:

  • Kill the tmux session
  • Remove the git worktree
  • Delete the git branch
  • Clean up all session metadata
# Via CLI
core-cli sessions delete task-42-abcde

# Via chat (AI will use the tool)
"Delete the session for task 42"

Clear All Sessions

# Via CLI
core-cli sessions clear

# Via chat
"Clear all coding sessions"

Example Workflow

Here's a complete example of delegating tasks:

# 1. Start core-cli
core-cli

# 2. View your tasks
/tasks

# 3. Delegate a task to Claude Code
"Implement task 5"

# Output:
# ⚑ launch_coding_session
# β”” Task: #5
#   Description: Add user authentication API
#   Agent: claude-code
#
# SUCCESS: Launched Claude Code in background session "task-5-xfgpq"
#
# To attach to this session, run: core-cli sessions attach task-5-xfgpq

# 4. Check active sessions
/sessions

# Output shows:
# πŸ’» Active Coding Sessions
#
# 1. [claude-code] task-5-xfgpq
#    Task: #5 - Add user authentication API
#    Status: detached | Started: 2m ago
#    Worktree: /Users/you/worktrees/task-5-xfgpq
#    Branch: task-5-xfgpq
#    Attach: core-cli sessions attach task-5-xfgpq

# 5. Attach to the session
core-cli sessions attach task-5-xfgpq

# 6. Work with Claude Code in the tmux session
# When done, press Ctrl+B then D to detach

# 7. Later, delete the session
core-cli sessions delete task-5-xfgpq

Advanced Configuration

Custom Agent Paths

If your coding agent is installed in a non-standard location, configure it in Core:

// In your Core CLI config
{
  "codingAgents": {
    "claude-code": {
      "path": "/custom/path/to/claude"
    }
  }
}

Requirements

  • tmux must be installed (brew install tmux on macOS)
  • git must be installed and the directory must be a git repository
  • The selected coding agent must be installed and available in PATH

Troubleshooting

"tmux is not installed"

# macOS
brew install tmux

# Ubuntu/Debian
apt-get install tmux

"Agent not found"

  • Ensure the coding agent is installed
  • Check that it's in your PATH: which claude or which cursor
  • Configure custom path if needed

"Not a git repository"

  • Initialize git: git init
  • Or run Core CLI from within a git repository

"Multiple sessions found for task"

  • Use the full session name: core-cli sessions attach task-42-abcde
  • Or delete old sessions first: core-cli sessions delete task-42-xxxxx

Development

Setup

# Clone the repository
git clone https://github.com/RedPlanetHQ/core-cli.git
cd core-cli

# Install dependencies
pnpm install

# Build the project
pnpm run build

# Run in development mode
pnpm run dev

Architecture

Core CLI is built on three key components:

  1. CLI Interface - React + Ink for beautiful terminal UI
  2. Core Memory API - Temporal knowledge graph with 88% recall accuracy
  3. MCP Integration Layer - Model Context Protocol for tool execution
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚           Core CLI (You)                β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
              β”‚
              β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚        AI Provider (LLM)                β”‚
β”‚    OpenAI / Anthropic / Custom          β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
              β”‚
              β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚      Core Memory (Knowledge Graph)       β”‚
β”‚  β€’ Temporal relationships                β”‚
β”‚  β€’ Cross-domain links                    β”‚
β”‚  β€’ Entity extraction                     β”‚
β”‚  β€’ Smart clustering                      β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
              β”‚
              β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚     MCP Servers (Integrations)          β”‚
β”‚  Gmail β€’ Calendar β€’ GitHub β€’ Linear     β”‚
β”‚  Docs β€’ Sheets β€’ Slack β€’ File System    β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

Use Cases

For Individual Developers

  • Never Lose Context - All your tasks, conversations, and decisions remembered across sessions
  • Unified Task Hub - Manage GitHub issues, Linear tickets, and personal tasks in one place
  • Smart Prioritization - AI helps you decide what to work on based on deadlines and context
  • Automatic Organization - Weekly task files with smart rollover keep you organized without effort
  • Integration Magic - Sync routine pulls updates from all your tools and creates actionable tasks
  • Document-Linked Tasks - Reference meeting notes, design docs, and specifications directly in tasks
  • CLI Speed - Beautiful GUI interface without leaving the terminal
  • Your Data - Complete ownership and control of your task history and knowledge graph

Roadmap

  • Advanced task analytics and insights
  • More integration connectors (Jira, Asana, Trello)
  • Custom routine builder
  • Task templates and recurring tasks
  • Self-hosted Core memory option
  • Mobile companion app
  • Advanced memory visualization
  • Custom MCP server marketplace

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/amazing-feature)
  3. Commit your changes (git commit -m 'Add some amazing feature')
  4. Push to the branch (git push origin feature/amazing-feature)
  5. Open a Pull Request

Community & Support

License

This project is licensed under the GNU Affero General Public License v3.0 with Commons Clause.

See the LICENSE file for full details.

Note: The Commons Clause means you can use, modify, and distribute this software freely, but you cannot sell the software or services whose value derives primarily from the functionality of this software.

Acknowledgments

This project is based on Nanocoder by Nano Collective and contributors.

We are deeply grateful to the Nano Collective team for creating the foundation and inspiration for this CLI tool. Their work has been instrumental in making this project possible.

Special thanks to:

  • Nano Collective for the original Nanocoder project
  • All contributors to the Nanocoder project
  • The open-source community
  • Early Core users and beta testers

Built with ❀️ by RedPlanetHQ

Core CLI - The AI that never forgets.

About

AI task manager

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •