When to Use

User wants to operate Cursor as a real coding environment instead of a vague AI editor: choose the right surface, shape repo context, set rules correctly, use cursor-agent, run Background Agents, review with Bugbot, or decide whether MCP and remote execution are acceptable.

Use this skill when the hard part is not "write code" but "make Cursor behave predictably" across editor chat, Agent or Ask flows, CLI runs, project rules, repo indexing, remote agents, GitHub integration, and privacy-sensitive work.

Architecture

Memory lives in ~/cursor/. If ~/cursor/ does not exist, run setup.md. See memory-template.md for structure.

~/cursor/
|-- memory.md            # Durable activation boundaries and workflow defaults
|-- repo-profiles.md     # Per-repo conventions, trust posture, and verification expectations
|-- rules-notes.md       # Project-rule strategy, legacy rule cleanup, and instruction hierarchy notes
|-- privacy.md           # Indexing, ignore, remote execution, and data-handling defaults
|-- remote-workflows.md  # Background Agent, Bugbot, and GitHub integration decisions
`-- incidents.md         # Repeated failures, wrong-surface runs, and recovery patterns