Glossary

Repository or user guidance file that gives Codex persistent instructions.

Authentication using an OpenAI API key.

Codex asking to allow a restricted action.

Snapshot of the frontmost app window sent to a Work or Codex task.

Model-based review of eligible approval requests before they proceed.

Desktop app with ChatGPT Work, Codex, and a separate Quick Chat window, including projects, file previews, scheduled tasks, and developer tools.

Opt-in feature that builds memories from recent screen context.

Configured container setup used for Codex cloud tasks.

OpenAI's coding agent for software development tasks.

Terminal client for running Codex interactively or in scripts.

CLI command for running Codex non-interactively from scripts or CI.

Programmatic interface for building Codex-powered workflows or integrations.

Summarizing older context so long-running work can continue.

Desktop capability that lets ChatGPT interact with other applications through the UI.

Local Codex configuration files.

Connector that lets Work or Codex access external services through a plugin.

Information Codex can use while working, such as files, prior messages, tool output, and instructions.

User-defined agent role with its own instructions and settings.

Set of Git file changes shown for inspection, comments, staging, or reverting.

Desktop app configuration that tells Codex how to set up worktrees for a project.

Non-interactive run that skips saving session state after it completes.

Permission profile rule that grants or denies read and write access to paths.

Mode where Codex runs without normal sandbox restrictions.

Moving a task and its work between Local and Worktree.

A lifecycle handler that runs when a Codex event matches, such as tool use, permission requests, or when a turn stops.

Contiguous section of a diff that can be staged, unstaged, or reverted independently.

Real-time web lookup for current information.

A Work or Codex task that runs on the user's machine.

Organization-controlled Codex defaults and restrictions.

Readable context exposed by an MCP server for Codex to inspect.

Action exposed by an MCP server that Codex can call during a task.

Locally stored context Codex can reuse across sessions.

Permission for commands or environments to reach the internet.

CLI mode for running Codex from scripts or CI.

A long-lived worktree kept as its own project.

Codex's proposed or tracked steps for completing a task.

Plugin metadata file that identifies a plugin and points to bundled skills, apps, MCP servers, hooks, and metadata.

Named configuration preset for Codex.

A group of related chats or tasks and shared sources, or a local folder used for file-based work.

Codex review of changes or feedback on a pull request.

Mode where Codex can inspect but not modify without approval.

Connection that lets you access Work or Codex tasks on another device through a connected host.

Desktop app view for inspecting diffs, comments, and Git changes.

Enforced boundary limiting what Codex commands can access or modify.

SDK shorthand for common sandbox policies such as read-only, workspace-write, or full access.

One execution of a scheduled task that may report findings or archive itself.

Work scheduled from an existing task that returns to the same task context on each scheduled run.

Script run before the agent starts to install dependencies or prepare tools.

Explicit or implicit activation of a skill.

Scheduled task whose runs each start a new task and report findings in Triage.

MCP server reached over HTTP, optionally with bearer token or OAuth authentication.

Workflow where Codex runs delegated agents in parallel and combines their results.

Technical conversation object exposed by Codex app-server APIs and protocol fields.

One exchange in a chat or task, usually a user prompt plus the agent's response and actions.

Pre-indexed search results Codex can use without live browsing.

Mode where Codex isolates changes in a separate Git worktree.