Stage 7 · Lesson 2 of 5
Run Modes and approvals in Cursor
's , in Settings, Agents, Approvals & Execution, decide which commands, MCP tools and fetches run without asking. Auto-review runs allowlisted calls, runs other shell commands in a sandbox when it can, and sends the rest to a that can , redirect or ask you.
Allowlist runs only the actions you listed. Run Everything runs every tool call with no prompts. Cursor says plainly that Auto-review is not a security boundary: the classifier can be wrong.
You can steer it with plain-English block_instructions in .cursor/permissions.json. Note that the Agent edits workspace files without approval, except configuration files, and changes save immediately, so Git is essential.
Common trapchoosing Run Everything on your own laptop because the prompts were annoying.
A question to check yourself, an AI practice session and 2 cheat sheet entries come when you add the roadmap to your account. Add it
Resources for this stage
Start with
- GuideFreeBuilding Effective Agents
Anthropic · Advanced
Anthropic's clear, free essay on how agents work and when simpler workflows are better
Keep handy
Also worth your time
Auto-review, Allowlist and Run Everything, permissions.json and sandboxing
- OWASP Top 10 for LLM ApplicationsGo deeper
Prompt injection and excessive agency explained by the security community