How does the self-healing feature work?

The --heal command automatically identifies and fixes deprecated API fields, missing event names, and common shell script errors in your hook configurations to ensure compatibility with the latest Claude Code versions.

Can I use LLMs to evaluate hook logic?

Yes, Hook Architect supports Prompt-Based Hooks, allowing you to use LLM evaluation instead of traditional shell scripts to determine if a task is complete or if a tool should proceed.

Does Hook Architect support MCP tools?

Yes, it includes specialized pattern matching for Model Context Protocol (MCP) tools using the mcp__server__tool naming convention.

What are Claude Code hooks?

Hooks are specialized triggers that allow you to intercept and modify Claude's behavior during specific events, such as before a tool runs, after it finishes, or when a session starts.

Hook Architect

Name: Hook Architect
Author: Lego4005

byLego4005

•

Productivity & Workflow

Simplifies the creation and maintenance of Claude Code hooks with guided discovery, automated self-healing, and API compatibility checks.

Hook Architect is a comprehensive utility for Claude Code users to build and manage robust automation workflows. It provides a guided interface to design specialized hooks for tool execution, session management, and subagent orchestration. By offering automated auditing—grading your hooks from A to F—and self-healing capabilities for deprecated APIs, it ensures your project's safety and productivity hooks remain functional and optimized as the Claude Code environment evolves. Whether you're implementing file-guardians for security or auto-formatters for efficiency, Hook Architect streamlines the entire lifecycle of your custom AI behavior.

Key Features

01Guided discovery workflow for designing custom codebase hooks

02Automated 'self-healing' to fix deprecated hook logic and formatting

037 GitHub stars

04Library of hook templates for safety, productivity, and intelligence

05Performance auditing and security grading (A-F) for existing hooks

06Support for component-scoped and prompt-based LLM hooks

Use Cases

01Securing codebases by blocking sensitive file access or dangerous shell commands

02Automating post-edit tasks like linting, formatting, or running test suites

03Injecting custom project context and environment variables at session start

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lego4005/foxit-marketplace hook-architect

For use in Claude.ai and ChatGPT

Download Skill