foodforthought-cli 0.2.0

CLI tool for FoodforThought robotics repository platform - manage robot skills and data

FoodforThought CLI

GitHub-like CLI tool for the FoodforThought robotics repository platform.

Installation

pip install foodforthought-cli

Or install from source:

cd foodforthought-cli
pip install -e .

Configuration

Set environment variables:

export ATE_API_URL="https://kindly.fyi/api"
export ATE_API_KEY="your-api-key-here"

Usage

Initialize a repository

ate init my-robot-skill -d "A skill for my robot" -v public

Clone a repository

ate clone <repository-id>

Create a commit

ate commit -m "Add new control algorithm"

Push to remote

ate push -b main

Deploy to robot

ate deploy unitree-r1

Commands

Repository Management

• ate init <name> - Initialize a new repository
• ate clone <repo-id> - Clone a repository
• ate commit -m <message> - Create a commit
• ate push [-b <branch>] - Push commits to remote

Skill Pipeline

• ate pull <skill-id> [--robot <robot>] [--format json|rlds|lerobot] [--output ./data] - Pull skill data for training
• ate upload <video-path> --robot <robot> --task <task> [--project <id>] - Upload demonstrations for labeling
• ate check-transfer --from <source-robot> --to <target-robot> [--skill <id>] - Check skill transfer compatibility
• ate labeling-status <job-id> - Check labeling job status

Deployment & Testing

• ate deploy <robot-type> - Deploy to a robot (e.g., unitree-r1)
• ate test [-e gazebo|mujoco|pybullet|webots] [-r robot] - Test skills in simulation
• ate benchmark [-t speed|accuracy|robustness|efficiency|all] - Run performance benchmarks
• ate adapt <source-robot> <target-robot> - Adapt skills between robots

Safety & Validation

• ate validate [-c collision|speed|workspace|force|all] - Validate safety and compliance
• ate stream [start|stop|status] [-s sensors...] - Stream sensor data

Cursor IDE Integration (MCP)

FoodforThought CLI supports Model Context Protocol (MCP) for integration with Cursor IDE.

Installation

1. Install MCP dependencies:

   pip install -r requirements-mcp.txt
   

2. Get your MCP configuration from the FoodforThought homepage

3. Place mcp.json in ~/.cursor/mcp.json (global) or .cursor/mcp.json (project)

4. Set your ATE_API_KEY environment variable

5. Restart Cursor

Available MCP Tools

• ate_init - Initialize a new repository
• ate_clone - Clone a repository
• ate_list_repositories - List available repositories
• ate_list_robots - List robot profiles
• ate_compatibility - Check skill compatibility between robots
• ate_adapt - Generate adaptation plans for skills

See the MCP documentation for more information.