hooks-builder

Configure Cursor agent hooks via hooks.json to observe, block, or modify actions.

8|Updated Feb 8, 2026
One-click install
npx skills add https://github.com/kensaurus/cursor-kenji --skill hooks-builder-kensaurus
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hooks-builder
Source: https://github.com/kensaurus/cursor-kenji/tree/main/skills/hooks-builder
Command: npx skills add https://github.com/kensaurus/cursor-kenji --skill hooks-builder-kensaurus

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Hooks provide a structured way to automate, observe, and control the Cursor agent loop, reducing manual interventions and enabling governance over behavior.

Core Features & Use Cases

  • Define events such as sessionStart, preToolUse, beforeShellExecution, beforeMCPExecution, and more to observe, block, or modify agent actions.
  • Configure global or project-scoped rules via hooks.json to automate actions like auto-formatting, security gates, audit logging, and PII scanning.
  • Practical use cases include validating tool usage before execution, guarding terminal commands, and auditing agent activity for compliance.

Quick Start

Create a hooks.json at the project root and add a simple hook script to begin automating and observing the agent loop.

Frequently Asked Questions about hooks-builder

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate and control the Cursor agent loop with hooks?

You can automate and control the Cursor agent loop by defining a hooks.json file at the project or global scope, using event triggers like sessionStart or preToolUse to observe, block, or modify agent actions.

What events can I hook into to observe or block Cursor agent actions?

Available hook events include sessionStart, preToolUse, beforeShellExecution, and beforeMCPExecution, allowing you to observe, block, or modify agent actions during development workflows.

How do I set up security gates and audit logging for Cursor agent tool usage?

Set up security gates and audit logging by configuring rules in a hooks.json file and adding optional scripts in a scripts/ directory to validate tool usage or audit agent activity before execution.

Can I apply Cursor hooks globally or do I need project-scoped configurations?

Cursor hooks can be applied at both global and project scopes, requiring a hooks.json configuration file to define deterministic rules for automating actions like PII scanning and command guarding.

How do I validate terminal commands before the Cursor agent executes them?

Validate terminal commands by configuring the beforeShellExecution event in your hooks.json, which guards terminal commands and blocks unauthorized shell executions before they run.