agents.md is read before doing any work

How codex discovers guidance:

The instruction chain:

1.Global scope:

in your Codex directory(default~/.codex): u can also use AGENTS.override.md,but only the first non-empty AGENTS.md is read

2.Project scope:

At the project root in each directory along the path,it checks for AGENTS.override.md ,then AGENTS.md,then any fallback names in project_doc_fallback_filenames. Codex includes at most one file per directory

3.Merge order:

The later ,the higher : because they appear later in the prompt

Create global guidance:

1.make sure the directory exist: 2.create with reusable preference 3.run codex anywhere to confirm it loads the file

if you want a temporary global ,use AGENTS.override.md

Layer project instructions:

1.In your repository root, add an AGENTS.md that covers basic setup 2.Add overrides in nested directories when specific teams need different rules. For example, inside services/payments/ create AGENTS.override.md 3.Start Codex from the payments directory:

Add code reviews rules:

For Codex code review in GitHub, add a ## Code Review Rules section to the AGENTS.md closest to the code the rules govern. Put repository-wide checks at the root and service-specific checks in a nested file.

Customize fallback filenames:

if u use a customize filename,add it to the fallback list then Codex can really read it as a instructions file: 1.edit your Codex configuration

1
2
3

# ~/.codex/config.toml
project_doc_fallback_filenames = ["TEAM_GUIDE.md", ".agents.md"]
project_doc_max_bytes = 65536

2.Restart Codex or run a new command to update configuration loads:

set preference and carry useful context across chats:

Choose a personality:

Friendly, pragmatic,or None

Add custom instructions:

In your global AGENTS.md

Carry context forward with memories:

Memories are separate from required project guidance. Keep instructions that must always apply in AGENTS.md or checked-in project documentation.

Add recent activity with Computer History

Review what Computer History includes before enabling it

Manage personalization:

pen Settings to update your personality, custom instructions, memories, and other available personalization controls. See ChatGPT desktop app settings for an overview of everyday preferences.

Helps to complete repeatable work with the right instructions,resources,and tools,They reduce the need to paste the same prompt ,template,requirements,or process into every chat.

Skill:

A skill packages instructions and supporting resources for a specific task or workflow.

Plugin:

A plugin is an installable bundle that can include skills, connectors, or both. Connectors are backed by Model Context Protocol (MCP) servers and can optionally include custom ChatGPT UI.

Use skills for repeatable work:

A skill is a reusable workflow that gives ChatGPT or Codex task-specific guidance.

A skill can combine:

• A name and description that help ChatGPT and Codex recognize when the skill applies.
• Workflow instructions that define the process and expected result.
• Supporting resources such as templates, examples, brand guidance, schemas, or connected tools.

Build skills:

You can start by turning a task you already repeat into a focused playbook for ChatGPT and Codex.

To build a useful skill:

1. Choose one focused task. Note what you normally start with, such as files, links, or notes, and what a finished result should look like.
2. Describe the workflow. In ChatGPT, start with @skill-creator; in Codex, use $skill-creator. Explain the goal, the steps to follow, the expected format, and anything the skill should always include or avoid. Add a template or a good example when you have one.
3. Review and try the draft. Check the instructions, test the skill with a realistic request, and refine it if the result misses a step or drifts from the format you want.
4. Install and reuse it. Once the skill is enabled, ChatGPT or Codex can use it for relevant requests, or you can select it explicitly. You can also share it with teammates when your workspace settings allow it.

Use plugins for tools and shared workflows:

A plugin can combine skills with connectors for services such as GitHub, Google Drive, or Slack, and can include MCP servers for additional tools and context.

Choose between a skill and a plugin:

Use a skill when you need reusable instructions for a focused task.

Use a plugin when you want an installable package that can combine instructions with connected services or other tools.

overview

Larger task,plz include these parts: Goal Context Output: format,length,level of detail Boundaries

Describe the result u need:

not detailed list,but needed details those change what AI should produce

Add useful context:

for example: necessary attach file; visual work: screenshot,image;and point out which area matters web search: Explicitly( 显式地 ) ask to use Web Search in the conversation. share resources: using a directory

Use connected sources:

name where it should look and what if should find

Using plugins:

plugins provide reusable instructions and connections to tools Ask for the result u need and let the active surface choose from the tools available to it

Personalize :

Put preferences that should apply across chats in Settings > Personalization as custom instructions.

Set boundaries that prevent real problems:

avoid creating extra work or taking an action u did not asked.

• Keep the approved dates and budget figures unchanged.
• Use only the supplied sources. Flag missing information instead of guessing.
• Keep recommendations within the stated budget.
• Prepare the message as a draft. Don’t send it. focus on the one or two boundaries that really matters

Improve the result with follow-up messages:

Your first prompt doesn’t need to be perfect. Review the result, then ask for the specific change you want.

Steering and queuing:

Steer add to the current run;Queue for the next run. In the CLI,using Enter to steer,and Tab to queue.

Put the pieces together:

include the four parts(maybe not all) above

Use voice dictation:

Ctrl+Shift+Tab

Prompting examples for chat:

start with the outcome u want and add details when it changes to the answer

Understand a topic

1
2

Explain how compound interest works for someone who has never invested.
Use one concrete example and define any financial terms you introduce.

Draft and refine writing

1
2

Draft a friendly email declining this invitation because I will be traveling.
Keep it under 120 words and leave the door open for a future event.

Compare options

1
2
3

Compare these two phone plans for one person who travels internationally twice
a year. Show the important differences in a table, then recommend one and explain
the tradeoff.

Make a practical plan

1
2

Plan five weekday dinners that take less than 30 minutes. Avoid peanuts, reuse
ingredients across meals, and finish with one consolidated shopping list.

Prompting codex:

Name the behavior u want,points to the relevant code or codebase or reproduction steps,preserves important constraints and says how to verify the change.

For a muti-step task,/plan when u want codex to investigate and propose an approach before editing. /goal to set a persistent goal CLI: using /mention and @ path autocomplete Codex runs local commands inside a sandbox that limits file and network access

Explain a codebase:

1
2
3
4
5
6

Explain how the request flows through the selected code.

Include:
- a short summary of the responsibilities of each module involved
- what data is validated and where
- one or two "gotchas" to watch for when changing this

Fix a bug:

Give Codex a reproduction recipe, plus the file(s) you suspect

Context notes: by u and by Codex itself Verification: Run the repro steps after fix if u have a standard check pipeline ,ask it to run it

Write a test:

Prompt with a function name: by the way,function name should be meaningful