Claude Hero | The Free Pair Programming Tool That Makes Claude Code 10x Faster

What Is Claude Hero?

Claude Hero is an external tool that monitors your entire development workflow in real time by utilizing two hook events in Claude Code: UserPromptSubmit and PostToolUse. It automatically inserts security checks, test reminders, and workflow suggestions, and adopts a "non-blocking design" that only halts processing for critical issues.

Since hooks are registered in the hooks section of ~/.claude/settings.json, the tool operates on a separate layer from official plugins. Its strength is that it can be installed on any version of Claude Code as long as the hooks feature is enabled.

The official website is at claudehero.dev, and the source code is available at github.com/ohmnow/claude-hero.

4 Operating Modes

Claude Hero offers 4 modes tailored to different skill levels and use cases. You can switch between them at any time during a session with the /claude-hero command.

Minimal — The Lightweight Security Guardian

This mode only notifies you of security alerts and makes no other interventions. It blocks only hardcoded API keys and dangerous git operations, minimizing its impact on your workflow. It is suited for senior developers who already follow best practices and want to use it as a "last line of defense."

Training — The Educational Mode That Accelerates Learning

This mode provides educational feedback by attaching detailed explanations and documentation links to each suggestion. Because it carefully explains "why a particular operation is dangerous," it is ideal for developers who are learning a specific framework or language. Understanding the reasons helps prevent repeating the same mistakes.

Guidance — The Balanced Default Mode

This mode strikes a balance between suggestions and autonomy. It inserts reminders to run tests and perform pre-commit checks at the right moments without disrupting your development flow. If you are using Claude Hero for the first time, this is the recommended mode to start with.

Orchestrator (Hero Mode) — Autonomous Project Management

This is the full-power mode that functions as a "10x pair programmer." It parses a PRD (Product Requirements Document) in Markdown format, autonomously breaks down tasks, and tracks progress by sprint, phase, and feature unit. It acts as a command center when driving large-scale feature development with Claude Code.

Security Detection Features

The main patterns that Claude Hero automatically detects and blocks are as follows.

Because it does not stop your work except for critical security alerts, it prevents loss of focus due to frequent interruptions. It analyzes conversation transcripts to understand context, and is designed to deliver only suggestions relevant to what you are working on at that moment.

Installation

There are two ways to install Claude Hero.

Global install from npm (recommended)

npm install -g claude-hero

Source build from GitHub

git clone https://github.com/ohmnow/claude-hero.git
cd claude-hero
npm install && npm run build
node scripts/install.js

With either method, the installation script automatically appends the necessary hook configuration to ~/.claude/settings.json. Restarting Claude Code will activate Claude Hero.

Claude Code Hook Configuration Details

The installation script handles the configuration automatically, but if you want to configure it manually or review the settings, check the hooks section of ~/.claude/settings.json.

{
  "hooks": {
    "UserPromptSubmit": [
      {
        "hooks": [
          {
            "type": "command",
            "command": "claude-hero check-prompt"
          }
        ]
      }
    ],
    "PostToolUse": [
      {
        "hooks": [
          {
            "type": "command",
            "command": "claude-hero post-tool"
          }
        ]
      }
    ]
  }
}

UserPromptSubmit runs every time a prompt is submitted, handling content checks before submission. PostToolUse runs after every tool use such as file editing or command execution, providing feedback for each operation. The two hooks working together cover the entire Claude Code cycle of prompt → tool execution → next prompt.

Project Configuration (.claude-hero.json)

By placing a .claude-hero.json file in your project root, you can configure settings on a per-repository basis. All team members can share the same configuration.

{
  "mode": "guidance",
  "rules": {
    "security": true,
    "testing": true,
    "commits": true,
    "workflow": true
  },
  "cooldown": 300
}

Use mode to specify one of the four modes. The rules field lets you individually enable or disable each rule category, allowing flexible customization to suit your team's needs. cooldown is the number of seconds to wait before the same suggestion is repeated. The default is 300 seconds (5 minutes), which prevents "suggestion fatigue" from the same suggestion arriving repeatedly.

PRD Parser and Hero Mode

The PRD parser is the core feature of Orchestrator mode. It reads a specification document in Markdown format and automatically tracks task progress in checkbox format.

## Phase 1: Authentication Foundation
- [ ] Implement user registration API
- [ ] Session management
- [x] DB schema definition (complete)

## Phase 2: Payment Features
- [ ] Receive Stripe Webhook
- [ ] Recurring billing processing

When you write specifications in this format, Claude Hero understands the completion status of tasks and suggests what to work on next based on context. It also supports sprint phase tracking, functioning as a project management tool when driving long-term projects with Claude Code. Loading a PRD begins with the /claude-hero hero command.

Automatically Detected Tech Stacks

Claude Hero automatically detects your project structure and makes suggestions tailored to your stack. Supported tech stacks include:

• React / Next.js
• Node.js / Express
• TypeScript
• Python
• Django / Flask / FastAPI
• Monorepo structures

For Python projects, you will receive reminders to run pytest, and for TypeScript projects, suggestions for type checking will increase — the tool adapts contextually based on your stack.

Differences from Official Plugins

Claude Hero is not included in the Claude Code official plugin marketplace. While official plugins are installed via the /plugin install command, Claude Hero is a third-party tool that uses the hooks feature in ~/.claude/settings.json.

Combining it with the official commit-commands plugin lets you build a workflow where Claude Hero security checks run before a commit, followed by automatic PR creation via commit-commands. There is no conflict with official plugins, and using them in a complementary manner is the most effective approach.

Summary

Claude Hero is a free, open-source tool that leverages Claude Code's hooks feature. It provides three functions — security detection, pair programming support, and project management — across four operating modes. Installation is completed with a single command, npm install -g claude-hero, and configuration only requires appending hooks to ~/.claude/settings.json.

Minimal mode stays out of the way as a security guardian with minimal intervention, while Orchestrator mode autonomously manages the entire project by referencing your PRD. By choosing and switching modes according to your team's skill level and needs, you can progressively deepen your use of Claude Code.

Sources

• Claude Hero Official Website (claudehero.dev)
• github.com/ohmnow/claude-hero — MIT License

Related Articles

• Best Claude Code Plugins | Full Category List and How to Choose
• What Is Claude Code — 9 Features That Work Even for Non-Engineers
• Claude Code CLAUDE.md Infrastructure — A Full Explanation of Every Configuration Option
• The Structure and Creation of the Claude Code Marketplace