Commands and Workflows - Entry Points for Skills and Automation

Codex commands include built-in controls and reusable workflow entry points. Separate them by purpose.

Four Entry Points

Entry point	Purpose	Example
Built-in command	Control a session or UI	/agent, /skills, /plugins
Prompt file	Share a standard request	PR review criteria
Skill	Reuse procedures, references, and scripts	$docs-content
Non-interactive execution	Run repeatable automation	codex exec, Codex GitHub Action

Built-in commands control Codex itself. Put project procedures in Skills and move CI workflows to non-interactive execution.

Design a reusable prompt

# Pull request review

Compare this branch with main.
Report bugs, security risks, regressions, and missing tests.
List findings first with file and line references.
Do not edit files.

A reusable prompt should define the goal, input scope, restrictions, output format, and completion condition. Keep repository-wide permanent rules in AGENTS.md instead of duplicating them.

codex exec and the GitHub Action

The Codex GitHub Action can invoke codex exec with a prompt file, model, sandbox, and output file. Before automation, decide whether the task is read-only or can write patches, what network or secrets are available, how failures affect the job, and whether reruns are safe.

When to Move a Workflow into a Skill

A short instruction can remain a prompt file. Move it into a Skill when it needs multiple steps, supporting references, helper scripts, or explicit trigger rules.

Next, Skills design patterns explains progressive loading of task-specific procedures.