Learning HubACFS Academy
Progress
0%
0 of 3333 remaining
  • 1
    Welcome & Overview5 min
  • 2
    Linux Navigation8 min
  • 3
    SSH & Persistence6 min
  • 4
    tmux Basics7 min
  • 5
    Git Essentials10 min
  • 6
    GitHub CLI8 min
  • 7
    Agent Commands10 min
    NOW
  • 8
    NTM Command Center8 min
  • 9
    NTM Prompt Palette6 min
  • 10
    The Flywheel Loop10 min
  • 11
    Keeping Updated4 min
  • 12
    UBS: Code Quality Guardrails8 min
  • 13
    Agent Mail Coordination10 min
  • 14
    CASS: Learning from History8 min
  • 15
    The Memory System8 min
  • 16
    Beads: Issue Tracking8 min
  • 17
    Safety Tools: SLB & CAAM6 min
  • 18
    The Art of Agent Direction12 min
  • 19
    Case Study: cass-memory15 min
  • 20
    Case Study: SLB12 min
  • 21
    RU: Multi-Repo Mastery10 min
  • 22
    DCG: Pre-Execution Safety8 min
  • 23
    Meta Skill: Local Skills10 min
  • 24
    SRPS: System Protection8 min
  • 25
    JFP: Prompt Library6 min
  • 26
    APR: Automated Plan Reviser8 min
  • 27
    PT: Process Triage6 min
  • 28
    XF: X Archive Search6 min
  • 29
    RCH: Remote Compilation8 min
  • 30
    WA: WezTerm Automata8 min
  • 31
    Brenner Bot: Research10 min
  • 32
    GIIL: Cloud Image Downloads6 min
  • 33
    S2P: Source to Prompt6 min
Back to Home
Back
7/33
Lesson 7
10 min

Agent Commands

Talk to Claude, Codex, and Gemini

New to ACFS?

Complete the setup wizard first to get the most from these lessons.

Go to Set Up Accounts
Goal

Login to your coding agents and understand the shortcuts.

The Three Agents

You have three powerful coding agents installed, each from a different AI company:

Claude CodeAnthropic
claudecc
Codex CLIOpenAI
codexcod
Gemini CLIGoogle
geminigmi

What The Aliases Do

The aliases are configured for maximum power (vibe mode):

cc(Claude Code)
NODE_OPTIONS="--max-old-space-size=32768" \
  claude --dangerously-skip-permissions
  • Extra memory for large projects
  • Background tasks enabled by default
  • No permission prompts
cod(Codex CLI)
codex --dangerously-bypass-approvals-and-sandbox
  • Bypass safety prompts
  • No approval/sandbox checks
gmi(Gemini CLI)
gemini --yolo
  • YOLO mode (no confirmations)

First Login

Each agent needs to be authenticated once:

Claude Code

$ claude auth login

Follow the browser link to authenticate with your Anthropic account.

Codex CLI

On a headless VPS, Codex requires special handling because its OAuth callback expects localhost:1455.

Option 1: Device Auth (Recommended)

  1. Enable "Device code login" in ChatGPT Settings → Security
  2. Then run the command below
$ codex login --device-auth

Option 2: SSH Tunnel

  1. On your laptop, create a tunnel
  2. Then run codex login on VPS
# On laptop:
$ ssh -L 1455:localhost:1455 ubuntu@YOUR_VPS_IP
# Then on VPS:
$ codex login

Option 3: Standard (if you have a browser)

$ codex login
OpenAI Has TWO Account Types
ChatGPT (Pro/Plus/Team)
  • • For Codex CLI, ChatGPT web
  • • Auth via OAuth (`codex login`)
  • • Get at chat.openai.com
API (pay-as-you-go)
  • • For OpenAI API, libraries
  • • Uses OPENAI_API_KEY env var
  • • Get at platform.openai.com

Codex CLI uses ChatGPT OAuth, not API keys. If you have an OPENAI_API_KEY, that's for the API—different system!

If login fails: Check ChatGPT Settings → Security → "API/Device access"

Gemini CLI

$ gemini

Follow the prompts to authenticate with your Google account.

Backup Your Credentials!

After logging in, immediately back up your credentials:

bash
caam backup claude my-main-account
caam backup codex my-main-account
caam backup gemini my-main-account

Now you can switch accounts later with:

bash
caam activate claude my-other-account
Pro Tip
This is incredibly useful when you hit rate limits! Switch to a backup account and keep working.

Test Your Agents

Try each one to verify they're working:

bash
cc "Hello! Please confirm you're working."
bash
cod "Hello! Please confirm you're working."
bash
gmi "Hello! Please confirm you're working."

Quick Tips

  • Start simple - Let agents do small tasks first
  • Be specific - Clear instructions get better results
  • Check the output - Agents can make mistakes
  • Use multiple agents - Different agents have different strengths

Practice This Now

Let's verify your agents are ready:

bash
1# Check which agents are installed
2$ which claude codex gemini
3
4# Check your agent credential backups
5$ caam ls
6
7# If you haven't logged in yet, start with Claude:
8$ claude auth login
Pro Tip
If you set up your accounts during the wizard (Step 7: Set Up Accounts), you already have the credentials ready—just run the login commands!

Ready to level up?

Mark complete to track your learning progress.

Previous
GitHub CLI
Next
NTM Command Center