Skip to content

/codex-login

Log GPT (Codex) in on this machine with a ChatGPT device login - shows a link and a one-time code. For remote or headless sessions (Claude Code on the web). No question is sent to GPT.

From plugin
deliberation
18015 skills15 commands1 MCP
> /plugin marketplace add antonbabenko/deliberation
> /plugin install deliberation@antonbabenko-deliberation

Command definition

codex-login.md
name: codex-login
description: Log GPT (Codex) in on this machine with a ChatGPT device login - shows a link and a one-time code. For remote or headless sessions (Claude Code on the web). No question is sent to GPT.
allowed-tools: mcp__deliberation__codex-login
timeout: 60000

Codex login (GPT)

Starts, or joins, the ChatGPT device login for GPT on this machine and shows its link and one-time code. Nothing is asked of GPT. Use it when GPT has no login here, typically in a Claude Code web session.

Workflow

1. Call the tool, with no arguments:

   mcp__deliberation__codex-login({})

2. Show the result's `message` to the user **as-is**: never paraphrased, never truncated, and never reflowed - the link and the code each stand on their own line so they are one copy. By `status`:

  • `pending` - the message carries the link and the code. Tell the user to open the

link, sign in to ChatGPT, and enter the code; GPT answers from their next GPT call.

  • `authenticated` - GPT has a credential here, or the login just landed through a host

dialog. The message also says what to do if an existing credential turns out to be spent (a copied `auth.json`): the next GPT call replaces it, or `codex logout` first.

  • `starting` - codex has not printed its code yet. Call the tool again (it joins the

same login), up to 3 times, before showing anything; only if it is still `starting`, show the message and ask the user to run `/deliberation:codex-login` again shortly.

  • `declined` - only from an older server; current ones report the refusal by ending that

login, and the next call offers a new code.

  • `failed` / `unavailable` - show the message; it names the cause (no codex CLI, GPT

disabled in config, a login that ended). 3. Stop. Do not run shell commands, do not read or check `~/.codex/auth.json`, and do not call any other deliberation tool in the same turn.

Rules

  • **The user's action, not an answer** - the link and code are for the user to act on.

Present them plainly; do not summarize them away.

  • **One login per session** - calling this again, or calling `/ask-gpt` meanwhile, joins

the same login and shows the same code until it expires (15 minutes).

  • **Never copy credentials** - the login is this machine's own. Never suggest copying

`auth.json` from another machine: a ChatGPT refresh token works once, so a shared copy breaks on the first refresh.

Read more
@antonbabenko
Ships withdeliberation

Get a second opinion from GPT, Gemini, and Grok in Claude Code, Codex, Cursor, Kiro, Antigravity, OpenCode, or any other MCP host - plus local models (Ollama, LM Studio) and 400+ more models through OpenRouter, including Qwen, Kimi, and DeepSeek.

Stars on GitHub

Momentum

stars in total
180
Star history of antonbabenko/deliberation on GitHub, a chart by star-history.com

Other commands on deliberation.

analyze
Auto-invoked

analyze

Command

Analyze recent runs - per-model latency, tokens, and verdict agreement - and suggest model/reasoning/fanout tuning. Advisory,…

ask-all
Auto-invoked

ask-all

Command

Ask GPT, Gemini, Grok, and any configured OpenRouter models in parallel for independent second opinions, then synthesize and…

ask-gemini
Auto-invoked

ask-gemini

Command

Get Gemini second opinion on a question or current work. Single-shot, advisory, no contamination. Model pinned per call.

ask-gpt
Auto-invoked

ask-gpt

Command

Get GPT (Codex) second opinion on a question or current work. Single-shot, advisory, no contamination.

ask-grok
Auto-invoked

ask-grok

Command

Get Grok (xAI) second opinion on a question or current work. Single-shot, advisory, no contamination.

More commands from other plugins.

build
Auto-invoked

build

Command

Implement tasks incrementally — build, test, verify, commit. Add "auto" to run the whole plan in one approved pass.

@addyosmani@addyosmaniInstall Command