# Cursor

A coding environment with autonomous agents and cloud execution.

Revision: 2026-09-13.4 · Review: unreviewed

Commissioned by a Nimbalyst maintainer; not an independent comparison.

## Applicability

Unpinned public documentation as of 2026-09-13. Documentation review only.

## Claims

- **yes · documented**: Runs fleets of agents in parallel. [Cursor official documentation](https://cursor.com/en-US) Scope: Public documentation on the capture date; not executed.
- **yes · documented**: Cloud agents use their own computers to build and test features. [Cursor official documentation](https://cursor.com/en-US) Scope: Public documentation on the capture date; not executed.
- **yes · documented**: Presents completed agent work for review. [Cursor official documentation](https://cursor.com/en-US) Scope: Public documentation on the capture date; not executed.
- **yes · documented**: Plan Mode scopes implementation work. [Cursor · docs](https://cursor.com/docs) Scope: Documentation claim for the cited product surface; not runtime-tested. See source locator and dossier limitations.
- **yes · documented**: Supports MCP customization. [Cursor · docs](https://cursor.com/docs) Scope: Documentation claim for the cited product surface; not runtime-tested. See source locator and dossier limitations.
- **yes · documented**: Supports persistent rules. [Cursor · docs](https://cursor.com/docs) Scope: Documentation claim for the cited product surface; not runtime-tested. See source locator and dossier limitations.
- **yes · documented**: Supports skills, plugins and subagents. [Cursor · docs](https://cursor.com/docs) Scope: Documentation claim for the cited product surface; not runtime-tested. See source locator and dossier limitations.
- **yes · documented**: Lists multiple selectable model providers. [Cursor · docs](https://cursor.com/docs) Scope: Documentation claim for the cited product surface; not runtime-tested. See source locator and dossier limitations.
- **yes · documented**: Queues follow-ups and supports steering an active turn. [Cursor · agent overview](https://cursor.com/docs/agent/overview) Scope: Client-specific controls; some Agents Window steering behavior is rolling out. Documentation only; not runtime-tested.
- **yes · documented**: Delegates tasks to subagents with separate context windows. [Cursor · subagents](https://cursor.com/docs/subagents) Scope: Editor, CLI and cloud-agent surfaces; parent receives subagent results. Documentation only; not runtime-tested.
- **yes · documented**: Sends selected elements and drawn annotations with code context. [Cursor · design mode](https://cursor.com/docs/agent/design-mode) Scope: Design Mode in the Agents Window browser; voice input, not a claim of spoken replies. Documentation only; not runtime-tested.
- **yes · documented**: Accepts narrated instructions while design agents run. [Cursor · design mode](https://cursor.com/docs/agent/design-mode) Scope: Design Mode in the Agents Window browser; voice input, not a claim of spoken replies. Documentation only; not runtime-tested.
- **yes · documented**: Continues directing a desktop-backed agent from iOS. [Cursor · mobile control](https://cursor.com/docs/cloud-agent/mobile) Scope: Cursor 3.9.8+ and eligible account with Cloud Agents access. Agent loop moves to cloud; tools stay local. Documentation only; not runtime-tested.

## Open questions

- Test reproducible environment setup and artifacts for cloud tasks.
- Differentiate model choice from support for external agent runtimes.

## Verification plan

1. Pin a product release or repository commit and record the host environment.
2. Create a disposable repository with a passing baseline test.
3. Run a bounded change with two agents where supported; record permissions and logs.
4. Record a failed attempt, actual outcome, and artifacts without secrets.
5. Have another person reproduce the result before labeling it reproduced.

Last checked: 2026-09-13. Recheck by: 2026-10-13.

Content rights: Unlicensed prototype; redistribution decision pending.