# Cursor ships Auto-review run mode for reduced-approval agent execution

> Cursor adds Auto-review, a new run mode that lets the agent run longer with fewer approval prompts. Allowlisted and sandboxable calls run immediately; other actions go to a classifier subagent that decides whether to allow, try a different approach, or ask for approval. Configurable in settings with optional custom classifier instructions.

| | |
|---|---|
| **Tool** | Cursor |
| **Version** | — |
| **Kind** | announcement |
| **Published** | 2026-05-29 |
| **Observed** | 2026-08-11 |
| **Significance** | 3/5 |
| **Breaking** | no |
| **Categories** | feature, capability |

## What changed


- Allowlisted Shell, MCP, and Fetch tool calls run immediately
- Sandboxable calls execute in the sandbox
- Unmatched agent actions route to a classifier subagent for allow/retry/approval
- Configuration: Settings > Cursor Settings > Agents > Approvals & Execution
- Classifier agent supports custom instructions


## Sources

- [blog_rss](https://cursor.com/changelog/auto-review) — retrieved 2026-08-11


## Community

_No curated reactions recorded. Facts and community takes are kept in separate layers
and never blended._

---
Canonical: https://changelogs.info/cursor/cursor-ships-auto-review-run-mode-for-reduced-approval-agent-execution
Entity: https://changelogs.info/cursor
Event ID: `evt_2026-05-29_cursor_auto-review-run-mode`
Licence: event synthesis © changelogs.info, CC BY 4.0. Linked sources belong to their vendors.