AgentHub

codex-cli

Drive the OpenAI Codex CLI (`codex`) reliably — both `codex review` (an independent, different-model-family second opin…

Live
Open / InstallLast updated July 27, 2026

Description

Drive the OpenAI Codex CLI (`codex`) reliably — both `codex review` (an independent, different-model-family second opinion on a diff/branch/PR/commit) and `codex exec` (run a coding task non-interactively: explore, refactor, run commands). Use this whenever the user wants to "review with codex", "get codex's opinion", a "second pair of eyes" / independent review before merging, OR to run `codex` / `codex exec` headlessly, pick its model, set reasoning effort, run it read-only, auto-approve its edits, or resume a codex session. Reach for it ESPECIALLY when a scripted `codex` call misbehaves — it documents and works around the flag traps agents keep hitting (exec hanging on stdin, no `-a`/`--full-auto` in exec, review's prompt-vs-scope conflict) and bundles a wrapper that just works. Not for Claude's own review (`/code-review`) or the Antigravity `agy` CLI (see the antigravity-cli skill). A reusable SKILL.md agent skill by Vl4dk0.

Author

Vl4dk0

Platform

cli

Pricing model

free

Categories

Skills

Tags

skill
claude-skill
github
en

Capabilities

  • SKILL.md package