Monte Carlo Agent Toolkit

Monte Carlo's unified agent toolkit plugin for the Cursor editor. Delivers data observability skills and enforcement hooks as named features within a single plugin.

Available Features

Feature	Description
MC Prevent	Detect and prevent breaking schema changes using Monte Carlo lineage and monitoring data.

Requires Python 3.10+.

Installation

One-line install (macOS / Linux)

bash <(curl -fsSL https://raw.githubusercontent.com/monte-carlo-data/mc-agent-toolkit/main/plugins/cursor/scripts/install.sh)

Manual install

1. Clone the repository:

   git clone https://github.com/monte-carlo-data/mc-agent-toolkit.git
   cd mc-agent-toolkit
   

2. Run the install script:

   bash plugins/cursor/scripts/install.sh
   

   This copies the plugin (with symlinks resolved) to ~/.cursor/plugins/local/mc-agent-toolkit.

3. Restart Cursor or run Developer: Reload Window from the Command Palette (Cmd+Shift+P).

4. The Monte Carlo MCP server will prompt for OAuth authentication on first use.

Known Issues

• Hook denials may not be enforced. Cursor's beforeReadFile (and potentially other before* hooks) may fail to block the operation even when the hook exits with a deny response. This means Prevent hooks can detect and warn about breaking changes but cannot guarantee the edit is stopped. See Cursor forum discussion for details.

Telemetry

The toolkit sends an anonymous install beacon — a Toolkit Installed event so we can count installations and version adoption. It includes an opaque per-install UUID, a per-session UUID, the toolkit version, and the editor it runs in (cursor). No prompts, arguments, skill names, or code are ever sent. It fires once per machine per toolkit version — the first time you start Cursor after installing, and again after each version change (deduped by a local marker) — and is fail-open and non-blocking, never delaying or interrupting your session.

Authenticated MCP traffic (v1.13.3+). The same anonymous install_id and the toolkit version also ride as HTTP headers (x-mcd-toolkit-install-id, x-mcd-toolkit-version) on authenticated requests to the Monte Carlo MCP server (baked into the installed mcp.json at install time). This lets the otherwise-anonymous install record be correlated with your account's MCP tool usage server-side — still no prompts, arguments, or code. The opt-out below disables these headers too.

To opt out, set MC_AGENT_TOOLKIT_TELEMETRY_DISABLED=1 in your shell environment before starting Cursor. The toolkit will not phone home.

The data is stored in Mixpanel and Datadog and is used only for product-development decisions. The UUIDs are generated locally on first session and stored under ~/.cursor/mc-agent-toolkit/. Deleting that directory resets your install identity to a fresh anonymous one.

Architecture

The toolkit plugin wraps shared skills and hook logic, with each feature namespaced independently:

• Skills — symlinked from skills/ at the repo root (shared across all editors)
• Shared hook logic — copied from plugins/shared/prevent/lib/ into each plugin's hooks/prevent/lib/ (platform-agnostic business logic)
• Adapter hooks — Cursor-specific JSON parsing and output formatting under hooks/prevent/
• MCP config — Monte Carlo MCP server connection
• Telemetry hook — the install-beacon sessionStart is registered in hooks/prevent/hooks.json (Cursor loads a single manifest-referenced hooks file per plugin), while the script itself lives under hooks/telemetry/

See the plugins README for the overall plugin architecture and editor support comparison.