codex-configuration

Configure Codex CLI config.toml sections, profiles, MCP servers, and approval policies.

28|1|Updated Aug 30, 2025
One-click install
npx skills add https://github.com/outfitter-dev/agents --skill codex-configuration
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex-configuration
Source: https://github.com/outfitter-dev/agents/tree/main/agent-kit/skills/codex-configuration
Command: npx skills add https://github.com/outfitter-dev/agents --skill codex-configuration

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Manages OpenAI Codex CLI configuration including config.toml settings, MCP servers, and model profiles.

Core Features & Use Cases

  • Profiles: Define named model configurations
  • Sandbox & approvals: manage sandbox modes and approval policies
  • MCP servers: integrate with MCP servers for automation

Quick Start

Edit your ~/.codex/config.toml or override with codex -c 'shell_environment_policy.inherit=all' to apply a quick change.

Frequently Asked Questions about codex-configuration

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I configure Codex CLI model profiles for different workflows?

Model profiles let you define named configurations in config.toml. Create profiles under the [profiles] section with model settings, then switch between them using codex -c or set your default profile. Each profile can specify different model parameters, sandbox modes, and approval policies tailored to your workflow.

Can I override Codex CLI settings at runtime without editing config.toml?

Yes, use the codex -c flag to apply runtime overrides. Pass configuration keys as arguments like codex -c 'shell_environment_policy.inherit=all' to temporarily change settings without modifying your config file, useful for testing or quick adjustments.

How do I set up MCP servers with Codex CLI?

Configure MCP servers in the [mcp_servers] section of config.toml. Specify server endpoints and their settings, then Codex integrates with them for automation tasks. Multiple servers can run concurrently across local, project, and repository scopes.

What's the difference between local, project, and repository configuration scopes?

Configuration scopes control precedence: repository settings override project settings, which override local settings. Skills paths follow strict precedence rules. Use repository scope for team standards, project scope for workflow specifics, and local scope for personal preferences.

How do I manage sandbox mode and approval policies in Codex CLI?

Configure sandbox_mode and approval policies in config.toml to control execution safety. Sandbox mode restricts file operations, while approval policies require consent before sensitive actions. Combine with exec_timeout_ms to set execution limits and control resource usage.

Why isn't my Codex CLI configuration applying as expected?

Check scope precedence—repository and project configs override local settings. Verify syntax in config.toml, confirm the file_opener and shell_environment_policy settings match your requirements, and use codex -c to test overrides. Review skills paths to ensure correct precedence order.