Do you know Cursor's less obvious features?
Five questions on subagents, Cloud Agents, .cursorignore, rule types and skills. You'll see each answer as soon as you pick. No account needed.
-
Q1 You're writing a custom subagent to review code. Which setting makes sense for it?Why
A review subagent shouldn't have write access. Each subagent runs in its own context and starts clean, which is why Cursor's docs suggest them for independent checks. Its
When Should You Use Subagents in Cursor?descriptionalso decides when Agent delegates to it. -
Q2 You're looking for "background agents" in Cursor and can't find them. Where are they?Why
A Cloud Agent runs on its own virtual machine, so you can close your laptop and review a pull request later. You can also start one at cursor.com/agents, or with
How Do You Hand a Task to a Cursor Cloud Agent?@cursorin Slack, GitHub or Linear. -
Q3 You add secrets.json to .cursorignore. What can still read it?Why
Listed files are blocked from Agent, Tab and
How Do You Keep Files Out of Cursor With .cursorignore?@mentions, but Cursor's docs say terminal and MCP tools can't be blocked by this file. So it isn't a security boundary. -
Q4 Your rule has alwaysApply: true and globs for src/api. Where does it load?Why
With
Which Cursor Rule Type Should You Use?alwaysApply: true, the rule is always included and the other two fields are ignored. To scope it by path, turnalwaysApplyoff and keep the globs. -
Q5 Your long deploy procedure is only needed once a week, and it's sitting in an always-on rule. What should you do?Why
Rules are for what the agent should always know, and skills are for what it can do when asked. Turn off model invocation for one with real side effects, such as deploying.
When Should a Cursor Rule Be a Skill Instead?