ChatGPT Dots Controls: Custom Rules and Approvals
Understand ChatGPT dots Activity View, Custom Rules, approvals, and sensitive actions that require you to take over, using OpenAI’s published guidance.
This guide restates only what OpenAI published on Introducing dots and How we build safety, security, and privacy into dots. It does not add click-paths.
Activity View
On How we build safety, security, and privacy into dots, OpenAI says Activity View in the desktop app shows ongoing and delegated tasks.
On that safety post, OpenAI says you can add context, correct a misunderstanding, change direction, or ask a dot to stop.
Custom Rules
On How we build safety, security, and privacy into dots, OpenAI describes Custom Rules that can allow an action, require approval, or block it.
On that safety post, OpenAI says Custom Rules cannot remove mandatory confirmations, handoffs, or core safety requirements.
On that safety post, OpenAI says dots can help write Custom Rules but need your approval to change them.
Introducing dots also describes Custom Rules.
Confirmations OpenAI says you still make
On How we build safety, security, and privacy into dots, OpenAI says some actions need confirmation each time: permanently deleting data, installing or running software from an unrecognized source, and granting new security-sensitive access.
On that safety post, OpenAI says purchases that use cards already saved on a merchant site require your approval.
Tasks that stay with you
On Introducing dots, OpenAI says some sensitive tasks, such as changing a password, always stay with the user.
On How we build safety, security, and privacy into dots, OpenAI says transferring money between financial accounts must be handed back to the user.
On that safety post, OpenAI says dots can help with the surrounding task.
When monitoring pauses work
On How we build safety, security, and privacy into dots, OpenAI says that if safety monitoring flags a concern in active work, it can pause that work and show a warning to review.
On Introducing dots, OpenAI also says monitoring can pause or stop the work.
Proactive research stays read-only
On Introducing dots, OpenAI says that when you are not actively working with it, the dot looks for ways to help using already-connected apps with read-only tools.
In that mode, Introducing dots says it cannot send messages, change app content, or control the browser or computer.
For the training exception OpenAI names around notes and proactive research, see Safety, sign-in, and data.
Usage limits OpenAI names
On Introducing dots, OpenAI says conversations with your dot do not count toward your ChatGPT usage limits.
On Introducing dots, OpenAI says tasks you ask the dot to start or manage in Codex or ChatGPT Work still count toward usage limits as usual.
For the deeper-work allowance, eligible plans, and rollout checks, see ChatGPT dots pricing, limits, and availability. To understand why a particular action needs approval, read how dots Auto-review works.
Permissions and approvals checked October 1, 2026
On Meet dots, OpenAI says: “Before an action affects your accounts or shares information, review determines whether your dot can proceed, needs your approval, or must hand a step over to you.”
On that Learn page, OpenAI says: “Neither instructions nor custom rules override built-in safety requirements.”
On Getting started with your dot, OpenAI says the Custom rules behaviors are “Take action without asking,” “Take action if pre-approved,” “Ask before taking action,” and “Hand off to you.”
On that Help Center page, OpenAI says: ““Pre-approved” means you explicitly requested the action in your prompt.”
On Manage dots in ChatGPT workspaces, OpenAI says Use custom rules for dots “Allows members to add or edit rules that guide their dot’s actions and confirmations. Off by default for Enterprise.”
On that workspace page, OpenAI says: “Custom rules cannot override built-in safeguards, and disabling custom rules does not make every action require approval.”
On Introducing dots, OpenAI says: “Dots use auto-review to check actions that could affect your accounts or share information against your instructions, Custom Rules, and safety requirements.”
On that announcement, OpenAI says: “Certain sensitive tasks, such as changing a password, always stay with you.”
A full list of every action that needs approval is 未公布 / not published on the pages opened for this update.
What this page does not add
This page does not invent approval menus, setup times, or sample prompts. For creating the first dot, see Get started with Dots.