/start-0-6.en
Lesson command
$ npx -y skills add minicoohei/ai-agent-camp --agent claude-codeHow it fires
How this command gets triggered: by you, by Claude, or both.
- Fires itselfClaude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/start-0-6.en
Context preview
What this command does when you run it.
Lesson command
Command definition
start-0-6.en.mddescription: "Lesson command"
chapter: "courses/aiagent/lesson02-setup/ch01-environment"
duration: "~15 min"
prerequisites: ["Node.js 18 or higher is installed", "OpenAI API key obtained"]
level: "beginner"
tags: ["setup", "codex", "cli"]
nonInteractiveMode: compliant
Lesson 0-6: Codex CLI Setup
Check Setup Progress
**Auto-run by AI:** Run `uv run python tools/setup_progress.py show` to display the current setup progress.
---
What You'll Do
| Item | Details | |------|---------| | Goal | Install and authenticate Codex CLI so you can run lessons in ai-agent-camp | | Duration | ~15 min | | Prerequisites | Node.js 18 or higher; OpenAI API key obtained | | Course Page | Refer to [Course Materials Top](https://ai-agent.camp/en/course/module-0) in parallel |
> **Hint**: This lesson is for Codex CLI users. Cursor users should start from Lesson 0-1.
---
Step 1: Install Codex CLI
Install Codex CLI via npm. Run the following in your terminal:
**Recommended: Run directly with npx (no installation required)**
npx @openai/codex --version
Using npx lets you run the latest version without a global install.
**Alternative: Global install**
If you use nvm or fnm, sudo is not required:
npm install -g @openai/codex
codex --version
> **Note**: Node.js 18 or higher is required. Check with `node --version`. > If you get a permissions error, refer to the [npm official guide](https://docs.npmjs.com/resolving-eacces-permissions-errors-when-installing-packages-globally) to change the prefix. `sudo npm install -g` is not recommended.
---
Step 2: Authentication (OpenAI API Key)
Codex CLI authenticates with an OpenAI API key. Configure it using one of the following methods:
Method A: Use credential_manager (Recommended)
Use `tools/credential_manager.py` to manage API keys securely:
uv run python tools/credential_manager.py store OPENAI_API_KEY
Follow the prompt to enter your key. It will be saved in an encrypted key store.
Method B: Set as environment variable
export OPENAI_API_KEY="your-api-key-here"
Add it to `.bashrc` or `.zshrc` to persist across sessions.
Method C: Set in .env file (fallback)
Only if the above methods are unavailable, add the following to the `.env` in the ai-agent-camp repository:
OPENAI_API_KEY=your-api-key-here
> **Security warning**: The `.env` file contains sensitive information. **Never commit it to Git.** Always verify that `.env` is included in `.gitignore`. Accidental commits risk API key leakage.
---
Step 3: Runtime Settings
Recommended settings for Codex CLI:
Codex CLI's approval mode is specified with `-a` (`--ask-for-approval`):
| Approval Mode | Description | |--------------|-------------| | `on-request` | The model automatically determines when to request user approval (recommended for learning) | | `never` | Auto-execute without confirmation (advanced users only, not recommended) |
Example startup command:
codex -a on-request
> **Important**: Codex manages the sandbox automatically. Follow the recommended settings in `AGENTS.md` for detailed configuration. Do not use `never` mode during normal learning. See `docs/codex-safety.md` for details.
---
Step 4: How to Run Lessons in Codex
In Cursor, you start lessons with slash commands like `/start-0-1`, but in Codex CLI you use **skills** instead.
Slash Command to Skill Mapping
| Cursor Command | Codex Method | |----------------|-------------| | `/overview` | Use the `aiagent-guide` skill | | `/check-setup` | Use the `aiagent-check-setup` skill | | `/start-0-1` | Use the `aiagent-lesson-runner` skill with `start-0-1` | | `/setup-security` | Use the `aiagent-tooling-setup` skill |
How to Use
Launch Codex CLI and make a request like:
Use the aiagent-lesson-runner skill to start the start-0-1 lesson
Or:
I want to start the start-0-1 lesson
Codex automatically recognizes `AGENTS.md` and the `skills/` directory and uses the appropriate skill.
---
Step 5: Verify Operation
Follow these steps to confirm Codex CLI is working correctly:
1. **Launch Codex in the ai-agent-camp directory**:
cd /path/to/ai-agent-camp
codex
2. **Verify the repository's hook configuration**:
Please run bash scripts/install_hooks.sh
3. **Run the setup check skill**:
Use the aiagent-check-setup skill to verify the environment
---
Expected Output Example
Environment Check Report
| Item | Status | Details |
|------------|--------|-----------------|
| Node.js | OK | 22.x |
| Codex CLI | OK | 1.x.x |
| OpenAI API | OK | Authenticated |
| Git | OK | 2.x |
| Hooks | OK | pre-commit configured |
Common Troubleshooting
- `codex: command not found` -> Run directly with `npx @openai/codex` or re-run `npm install -g @openai/codex`
- API authentication error -> Verify `OPENAI_API_KEY` is correctly configured
- Permission error -> Use nvm/fnm or [change the npm prefix](https://docs.npmjs.com/resolving-eacces-permissions-errors-when-installing-packages-globally)
- Skill not found -> Verify you launched Codex from the ai-agent-camp root directory
---
Checkpoint
- [ ] Codex CLI is installed (`codex --version` works)
- [ ] OpenAI API key is configured
- [ ] Approval mode is set to `on-request` (see recommended settings in AGENTS.md)
- [ ] Hooks are configured with `bash scripts/install_hooks.sh`
- [ ] The `aiagent-check-setup` skill runs successfully
---
Next Steps
Once Codex CLI setup is complete, you can start lessons.
**Recommended flow for Codex users:**
1. Setup verification: Run `start-0-1` (Environment Setup Verification) with the `aiagent-lesson-runner` skill 2. Start lessons: Begin `start-1-1` (Module 1 Banner Generation Intro) with the `aiagent-lesson-runner` skill 3. Execute each lesson's s
Read more
description: "Lesson command" chapter: "courses/aiagent/lesson02-setup/ch01-environment" duration: "~15 min" prerequisites: ["Node.js 18 or higher is installed", "OpenAI API key obtained"] level: "beginner" tags: ["setup", "codex", "cli"] nonInteractiveMode: compliant
Lesson 0-6: Codex CLI Setup
Check Setup Progress
**Auto-run by AI:** Run `uv run python tools/setup_progress.py show` to display the current setup progress.
---
What You'll Do
| Item | Details | |------|---------| | Goal | Install and authenticate Codex CLI so you can run lessons in ai-agent-camp | | Duration | ~15 min | | Prerequisites | Node.js 18 or higher; OpenAI API key obtained | | Course Page | Refer to [Course Materials Top](https://ai-agent.camp/en/course/module-0) in parallel |
> **Hint**: This lesson is for Codex CLI users. Cursor users should start from Lesson 0-1.
---
Step 1: Install Codex CLI
Install Codex CLI via npm. Run the following in your terminal:
**Recommended: Run directly with npx (no installation required)**
npx @openai/codex --version
Using npx lets you run the latest version without a global install.
**Alternative: Global install**
If you use nvm or fnm, sudo is not required:
npm install -g @openai/codex codex --version
> **Note**: Node.js 18 or higher is required. Check with `node --version`. > If you get a permissions error, refer to the [npm official guide](https://docs.npmjs.com/resolving-eacces-permissions-errors-when-installing-packages-globally) to change the prefix. `sudo npm install -g` is not recommended.
---
Step 2: Authentication (OpenAI API Key)
Codex CLI authenticates with an OpenAI API key. Configure it using one of the following methods:
Method A: Use credential_manager (Recommended)
Use `tools/credential_manager.py` to manage API keys securely:
uv run python tools/credential_manager.py store OPENAI_API_KEY
Follow the prompt to enter your key. It will be saved in an encrypted key store.
Method B: Set as environment variable
export OPENAI_API_KEY="your-api-key-here"
Add it to `.bashrc` or `.zshrc` to persist across sessions.
Method C: Set in .env file (fallback)
Only if the above methods are unavailable, add the following to the `.env` in the ai-agent-camp repository:
OPENAI_API_KEY=your-api-key-here
> **Security warning**: The `.env` file contains sensitive information. **Never commit it to Git.** Always verify that `.env` is included in `.gitignore`. Accidental commits risk API key leakage.
---
Step 3: Runtime Settings
Recommended settings for Codex CLI:
Codex CLI's approval mode is specified with `-a` (`--ask-for-approval`):
| Approval Mode | Description | |--------------|-------------| | `on-request` | The model automatically determines when to request user approval (recommended for learning) | | `never` | Auto-execute without confirmation (advanced users only, not recommended) |
Example startup command:
codex -a on-request
> **Important**: Codex manages the sandbox automatically. Follow the recommended settings in `AGENTS.md` for detailed configuration. Do not use `never` mode during normal learning. See `docs/codex-safety.md` for details.
---
Step 4: How to Run Lessons in Codex
In Cursor, you start lessons with slash commands like `/start-0-1`, but in Codex CLI you use **skills** instead.
Slash Command to Skill Mapping
| Cursor Command | Codex Method | |----------------|-------------| | `/overview` | Use the `aiagent-guide` skill | | `/check-setup` | Use the `aiagent-check-setup` skill | | `/start-0-1` | Use the `aiagent-lesson-runner` skill with `start-0-1` | | `/setup-security` | Use the `aiagent-tooling-setup` skill |
How to Use
Launch Codex CLI and make a request like:
Use the aiagent-lesson-runner skill to start the start-0-1 lesson
Or:
I want to start the start-0-1 lesson
Codex automatically recognizes `AGENTS.md` and the `skills/` directory and uses the appropriate skill.
---
Step 5: Verify Operation
Follow these steps to confirm Codex CLI is working correctly:
1. **Launch Codex in the ai-agent-camp directory**:
cd /path/to/ai-agent-camp codex
2. **Verify the repository's hook configuration**:
Please run bash scripts/install_hooks.sh
3. **Run the setup check skill**:
Use the aiagent-check-setup skill to verify the environment
---
Expected Output Example
Environment Check Report | Item | Status | Details | |------------|--------|-----------------| | Node.js | OK | 22.x | | Codex CLI | OK | 1.x.x | | OpenAI API | OK | Authenticated | | Git | OK | 2.x | | Hooks | OK | pre-commit configured |
Common Troubleshooting
- `codex: command not found` -> Run directly with `npx @openai/codex` or re-run `npm install -g @openai/codex`
- API authentication error -> Verify `OPENAI_API_KEY` is correctly configured
- Permission error -> Use nvm/fnm or [change the npm prefix](https://docs.npmjs.com/resolving-eacces-permissions-errors-when-installing-packages-globally)
- Skill not found -> Verify you launched Codex from the ai-agent-camp root directory
---
Checkpoint
- [ ] Codex CLI is installed (`codex --version` works)
- [ ] OpenAI API key is configured
- [ ] Approval mode is set to `on-request` (see recommended settings in AGENTS.md)
- [ ] Hooks are configured with `bash scripts/install_hooks.sh`
- [ ] The `aiagent-check-setup` skill runs successfully
---
Next Steps
Once Codex CLI setup is complete, you can start lessons.
**Recommended flow for Codex users:**
1. Setup verification: Run `start-0-1` (Environment Setup Verification) with the `aiagent-lesson-runner` skill 2. Start lessons: Begin `start-1-1` (Module 1 Banner Generation Intro) with the `aiagent-lesson-runner` skill 3. Execute each lesson's s
AI Agent Training for Non-Engineers - Complete Guide to Claude Code / Cursor / Codex ### ⚠️ Before you clone Official repository (maintained by the authors): Running AI agents from this repo grants them shell, file-write, and external-API permissions on your
Other commands on ai-agent-camp.
- /check-setup.en
Top-level alias — see lesson/check-setup.en.md for the full body.
Open command - /check-setup.es
Alias de nivel superior — el cuerpo completo está en lesson/check-setup.es.md.
Open command - /check-setup
Top-level alias — see lesson/check-setup.md for the full body.
Open command - /check-security.en
Lesson command
Open command - /check-security.es
Lesson command
Open command - /check-security
Lesson command
Open command

