OpenAI Codex Guide: Code Generation & Practical Tips

If you put several sites in one repository and place a CLAUDE.md or AGENTS.md in each site's folder, those rules aren't guaranteed to be read before work begins. Claude Code loads a CLAUDE.md below the launch directory when Claude reads, writes or edits a file in that folder, while Codex reads only the AGENTS.md files from the project root down to the launch directory, once at the start of each run. This article compares, item by item, how both tools load these files and how to check whether they were loaded (/context shows only what was loaded at startup), then covers a check that stops the release if anything outside the assigned folder changes, and tool settings that block writes. It also shows how to tell similarly named repositories apart by remote URL and root commit, and commands that show only environment variable names (checked November 1, 2026).

"Codex won't go parallel unless you tell it to" is true, with conditions: the Codex app, CLI and IDE extension spawn subagents when you ask directly, or when AGENTS.md or a skill instructs delegation (Ultra in ChatGPT Work also delegates on its own judgment). Claude Code delegates on its own based on the request and each subagent's description. Using the official OpenAI and Anthropic documentation (checked October 26, 2026), this article compares what triggers parallel work, five things to put in your instructions, ready-to-use example prompts, [agents] in Codex's config.toml and the TOML files in .codex/agents/, Claude Code's .claude/agents/ and environment variables, commands such as /subagents, /tasks, /subtask, /batch and /worktree, tasks that suit parallel work and tasks that don't, and the impact on usage limits, and it also shares lessons this site learned from running subagents in parallel.

This guide covers the causes of Codex feeling slow or heavy and how to fix them, organized by symptom, based mainly on OpenAI's official documentation and replies from GitHub maintainers as of October 25, 2026. For slow replies, check outages and congestion, lower reasoning effort to fit the task, use a lighter model or Fast mode (Fast uses 2.5 times the usage quota), and start a new thread for each task, as recommended by OpenAI staff. For slow startup, stop unused MCP servers with enabled = false and update the CLI, while a heavy or frozen app calls for an update and restart (one Windows bug was fixed in v26.825.32147) and archiving scheduled tasks you no longer need, since they pile up worktrees; on WSL, keep files on the Linux side to speed up commands.

This guide explains how to handle the Codex desktop app error "Organization settings could not be loaded" when it blocks you from continuing. The wording suggests a network or proxy problem, but GitHub bug reports show other causes too. It covers a case fixed by a single retry, four patterns reported by users, where to find the app logs, and what to avoid, such as signing out right away or deleting folders. It also explains what to report if the problem persists and how it differs from the similar "Couldn't load workspace settings" error.

We placed the app and its main data on drive D and kept active runtimes on drive C. Here is what we backed up, moved, and verified.

The page was visible, but Codex could not control it. We distinguish a missing connection from a stale error page and explain what actually recovered.

ChatGPT moved successfully, but destination preparation caused I/O failures in Docker and Unity on the same drive. Here is what the logs showed.

Covers setting up investigation-only use in the Codex app, IDE, and CLI, and in Claude Desktop, VS Code, web, and mobile. Explains read-only versus Plan, where settings apply and persist, tool restrictions, ready-to-use prompts, and diff checks.

Compares read denial, workspace separation, approvals, and network limits as ways to keep secrets away from Codex. Covers Permission profiles examples, .env naming and scope, environment variables, and OS-specific limits. Also explains how to confirm denial with a dummy file.

Compare Codex Remote and Cloud by where work runs, PC sleep behavior, phone access, development environments, and usage allowances. Learn the conditions for the connected PC, Cloud setup and unsupported features, Handoff limits, and how Work Cloud and dot differ. Also see what to check when a connection fails.

Use Codex /goal to keep a completion condition across investigation, fixes, and verification. This guide separates desktop and CLI operations, gives example requests, and covers pausing, resuming, clearing, and reasons work stops partway. It also explains how token budgets differ from pricing and what evidence to check in a completion report.

Before choosing Codex Security, compare the local plugin, the GitHub-connected Cloud service, and the standalone CLI. This guide uses official documentation to explain separate Cloud billing, conditional free scan credits, first-scan setup, Standard and Deep costs, scan coverage and reproduction evidence, and post-fix verification. It also covers when the standalone CLI uses an API key from the environment and how estimated cost limits work.