AI Power Progress iA
Interactive terminal learning

CLI Tutor

Practice Linux + Git + programming + math/compute + networking engineering. This terminal runs in a safe sandbox and can also use the AI coach. Use it to scaffold projects and hand off work via the Services Desk.

Loading CLI…

Terminal

$

Common Commands

This sandbox supports a focused set of commands for learning.

Ask the tutor

Use terminal commands:

Codex CLI (terminal)

Install and run Codex CLI in your Ubuntu terminal (not inside the web sandbox):
npm install -g @openai/codex
codex --login
codex

Local operator CLI (apctl)

Run these in your Ubuntu terminal (not inside the web sandbox). This is the fastest way to index docs, build the Daily Intelligence digest, run evals, and export feedback datasets.
cd ~/aipowerprogressia.com
chmod +x scripts/apctl

# health + model routing overview
./scripts/apctl health

# docs-only RAG (local)
./scripts/apctl index docs --dry-run
./scripts/apctl index docs
./scripts/apctl rag ask "What is this repo and how do I run it?"

# Daily Intelligence (Brave News + RSS + arXiv + CVEs)
./scripts/apctl news fetch
./scripts/apctl news digest --date today

# evals + feedback export
./scripts/apctl eval run
./scripts/apctl export feedback --limit 2000

Compute + Networking

Sandbox commands (local/token by default):

Project scaffolds

Generate starter code in the sandbox:

Ship SaaS services

Create a Services Desk request (works on the public site). Avoid sending secrets.

Other coding copilots

Run these in your real terminal (not inside the web sandbox):
# GitHub Copilot CLI (via GitHub CLI)
gh --version
gh extension list

# Claude / other CLIs
claude --version