2 min lesson
Tab, Agent and Ask are the trio, but the surface map is much bigger now
Imagine this comes up at work. "A new user tells you they avoid Ask mode because 'it runs a weaker model.' What's the accurate correction?" Start with the practical move.
Step 1 of 2
Tab, Agent and Ask are the trio, but the surface map is much bigger nowdon't get caught describing last year's Cursor
The Tab/Agent/Ask story is correct but dated if you stop there. A year ago Cursor's core was just three things: Tab, codebase indexing and chat. It now ships constantly. The most current framing from the field team is that Cursor is an end-to-end SDLC control plane: the same agent loop reaches across desktop, the agent window, web, mobile, the terminal, JetBrains and the cloud. An SA who still pitches "AI VS Code" is selling the 2024 product.
Cursor 3An older release label for Cursor's multi-agent interface; current product guidance calls the active surface the Agents Window. Press Enter for the full definition., internally called Glass or "the agent's window," is an agent-forward interface redesigned from the ground up to manage many parallel agents rather than edit one file. It sits one abstraction level above the editor: it shows every workspace across local and cloud, multiple repos and archived or active agents at once.
The driving rationale is the line to quote: as models improved, the team "needed to focus less and less on the output of the model and more and more on the outcome." You can still toggle back to the classic editor view, with the explorer left, code center, terminal bottom and agent pane right. Code and terminal are never hidden.
The other correction to make crisp: the mode dropdown changes the toolset, not the model. Picking Agent, Plan, Debug or Ask gives the same underlying model a different set of tools. New users assume "Ask modeA read-only mode for asking questions about a codebase without changing files; the safe way to explore unfamiliar or legacy code. Press Enter for the full definition. = a weaker model"; it's the same model with a read-only toolset. Get this right and you'll save the room a recurring confusion.
Learn more
Advanced table
Same model in each row; only the tools differ
- Mode
- Agent (default)
- What the toolset can do
- Everything: write code, run the terminal, make plans and to-dos
- When it wins
- A real task with a verifiable end state
- Mode
- Plan
- What the toolset can do
- Researches the whole indexed codebase and outputs only an editable plan; no edits until you switch to Agent
- When it wins
- Scoping a non-trivial change before any code is written
- Mode
- Debug
- What the toolset can do
- Walks you through reproduction, adds instrumentation, watches terminal and debug logs
- When it wins
- Diagnosing a symptom like "my search bar isn't working." BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. is the separate PR-review product
- Mode
- Ask
- What the toolset can do
- Read-only Q&A over the codebase; output stays in chat
- When it wins
- Onboarding to a new or legacy repo; non-technical users
| Mode | What the toolset can do | When it wins |
|---|---|---|
| Agent (default) | Everything: write code, run the terminal, make plans and to-dos | A real task with a verifiable end state |
| Plan | Researches the whole indexed codebase and outputs only an editable plan; no edits until you switch to Agent | Scoping a non-trivial change before any code is written |
| Debug | Walks you through reproduction, adds instrumentation, watches terminal and debug logs | Diagnosing a symptom like "my search bar isn't working." BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. is the separate PR-review product |
| Ask | Read-only Q&A over the codebase; output stays in chat | Onboarding to a new or legacy repo; non-technical users |
Same model in each row; only the tools differ.