cursor-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| CURSOR_API_KEY | Yes | User API key, required for all operations | |
| CURSOR_MCP_HOST | No | HTTP bind address | 127.0.0.1 |
| CURSOR_MCP_PORT | No | HTTP port | 11000 |
| CURSOR_INBOX_DIR | No | Message drop directory | ~/.cursor-mcp/inbox/ |
| CURSOR_USAGE_EMAIL | No | Filter spend by email | |
| CURSOR_ADMIN_API_KEY | No | Team admin API key, optional but required for spend/events endpoints | |
| CURSOR_RUNNING_AGENTS_WARN | No | Parallel agents warning count | 3 |
| CURSOR_ON_DEMAND_WARN_CENTS | No | On-demand spend warning threshold in cents | 2000 |
| CURSOR_HOURLY_SPEND_WARN_CENTS | No | Hourly spend warning threshold in cents | 300 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| logging | {} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| extensions | {
"io.modelcontextprotocol/ui": {}
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| cursor_usageA | Cursor billing and spend guardrails (read-only). Requires CURSOR_API_KEY. Admin endpoints also accept CURSOR_ADMIN_API_KEY (crsr_ team key). Use alert_check on a schedule (e.g. Fritz every 2h) instead of manual dashboard checks. |
| cursor_cloudA | List and inspect Cursor cloud agents (read-first; cancel is gated). Runaway spend often correlates with many parallel cloud agents — list often. |
| cursor_docsB | Return fleet-curated Cursor documentation snippets. |
| cursor_sdkA | Fleet-curated @cursor/sdk / cursor-sdk guidance (read-only, no agent spawn). Jun 2026 SDK release: custom tools, auto-review, JSONL stores, nested subagents. Use before writing new headless automation — complements cursor_docs topics. |
| cursor_inboxA | Inter-agent message drop for Cursor agents. Any process writes a message to the shared inbox dir (CURSOR_INBOX_DIR). The Cursor agent calls cursor_inbox to poll, read, and acknowledge. Typical flow:
No daemon, no network — pure filesystem. Default dir: ~/.cursor-mcp/inbox/ Override with CURSOR_INBOX_DIR env var. |
| cursor_helpA | List cursor-mcp tools and setup. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 6 tools
Each tool targets a distinct domain: cloud agents, documentation, help, inter-agent inbox, SDK guidance, and billing. No overlap in purpose, clear boundaries.
All tools follow a consistent 'cursor_<noun>' pattern with lowercase snake_case, making them easily predictable and grouped.
Six tools is well-scoped for the server's purpose, covering key Cursor ecosystem areas without being too sparse or bloated.
Covers cloud agent inspection, documentation, help, inbox messaging, SDK guidance, and billing. Minor gap: cloud agent management lacks create/delete, but this is likely intentional.