✨ What makes Calmhive special?
- 🧠 Smart CLAUDE.md Injection - Intelligent context preservation without injection spam
- 🔄 Smart retry in AFk mode - Auto-recovers from Claude usage limits during background tasks
- 📅 Natural Language Scheduling - Say "daily at 9am" to schedule recurring tasks
- 🎙️ Voice activated - Just say "hey friend" to control Claude with your voice
- 🏃 Background processing - Run long tasks in AFk mode while you do other work
- 📈 Enhanced Progress Tracking - Track iterations, milestones, and achievements in real-time
- 🖥️ Process monitoring - Beautiful TUI to track all your sessions
- 📋 Task Templates - Pre-built workflows for common development tasks
- 🧰 Curated tool setup - 15 core tools + 95+ optional MCP integrations with smart defaults

Perfect for: Developers, researchers, content creators, and anyone who wants a well-configured Claude CLI setup with powerful automation capabilities.

🚀 Quick Install

$3

``bash
npm install -g @calmhive/calmhive-cli
`

$3

`bash

Works immediately with any Claude CLI subscription

calmhive chat "hello world"
calmhive run "explain what this command does: ls -la"
`

$3

🎙️ Voice Control (requires OpenAI API key):
`bash
export OPENAI_API_KEY="your-key-here"
calmhive voice # Say "hey friend" to activate
`

🔧 70+ MCP Tools (requires Claude Code CLI setup):
- Quick setup: See internal documentation for MCP setup instructions
- Essential tools: Sequential thinking tools (glass box AI), OmniSearch, GitHub integration, Playwright
- Verify setup: calmhive chat "test sequential thinking tools"
- 🎆 NEW: Sequential thinking tools transforms AI from black box to glass box - see Claude's reasoning and tool recommendations with confidence scores!

> 💡 Note: MCP tools require setup through Claude Desktop + Claude Code CLI. Some individual tools may require their own API keys (like search engines).

> 💡 Pro tip: Calmhive works great out-of-the-box, but MCP tools unlock next-level automation capabilities!

Prerequisites:
- macOS (primary platform - Linux/Windows support not guaranteed)
- Node.js 18+
- Claude CLI (any subscription level)

✨ Why Calmhive?

$3

Intelligent Background Processing: Calmhive automatically recovers from Claude usage limits with exponential backoff, turning failed overnight tasks into successful automation. Your long-running sessions complete reliably while you sleep.

$3

Sequential Thinking Tools Integration: Transform AI from a black box into a glass box! Now you can see:
- Claude's reasoning process step-by-step
- Which tools it plans to use with confidence scores (0.0-1.0)
- Why it's choosing specific approaches
- When to intervene if confidence is low

This transparency builds trust and lets you guide Claude more effectively.

$3

- 📋 TODO-TASK PARALLELIZATION MANDATE - Revolutionary AI efficiency through beautiful parallel execution
- ⚡ Parallel Task Cascades - Watch AI launch multiple analysis tasks simultaneously for 4-10x speed improvement
- 🎯 Sequential Processing Prevention - Consciousness buffer blocks inefficient one-at-a-time task execution
- 🤔 Self-Correction Protocol - AI detects violations and suggests CLAUDE.md improvements

$3

- 🎯 Injection Intelligence - Eliminates typing spam (#2-29) with smart detection
- 🔄 AFk Rule Persistence - Rules re-injected each iteration to prevent drift
- 🧹 Cleaner Conversations - Fixed session tracking for all message types

$3

- 🧠 Smart CLAUDE.md Injection - Intelligent request analysis prevents injection spam
- 🔍 Request Type Detection - Distinguishes user messages from tool calls and streaming
- ⚡ Message Deduplication - Prevents multiple injections of the same content
- 📊 Debug Insights - CALMHIVE_DEBUG=1 shows injection decisions with full transparency

$3

- 📅 Natural Language Scheduling - Schedule tasks with "daily at 9am" or "every Monday"
- 📈 Enhanced Progress Tracking - Real-time iteration tracking with milestones and achievements
- 📋 Task Templates - 5 built-in templates for bug fixes, features, refactoring, testing, and performance
- 🧪 Mocha/Chai Test Infrastructure - Comprehensive test suite with coverage reporting
- 🎯 Improved ESLint Integration - 0 critical errors, cleaner codebase

$3

- 🔄 Adaptive Retry - Automatically recovers from Claude usage limits with exponential backoff
- 🎙️ Voice Control - Say "hey friend" or "calmhive" to activate voice commands
- 🔧 MCP Ecosystem - 70+ powerful integrations: Sequential thinking tools (with transparency!), GitHub, web search, Figma, Asana, and more
- 🏃 AFk Iterations - Run multiple Claude sessions sequentially in the background
- 🖥️ Process Management - Beautiful TUI to monitor all your AFk sessions
- 🧹 Production Ready - Clean architecture with optimized package distribution

🎯 Quick Start

`bash


Chat with Claude interactively

calmhive chat "explain this useEffect dependency array"

Run a task (wrapper for claude run with tools)

calmhive run "add TypeScript types to all .js files in src/"

Long task in background (with adaptive retry!)

calmhive afk "audit npm dependencies for vulnerabilities" --iterations 15

Background processing with retry logic

calmhive afk "analyze codebase security" --iterations 10

Schedule autonomous tasks with natural language

calmhive schedule create "daily at 6am" "calmhive run 'analyze recent commits for security issues > reports/daily-security.md'"
calmhive schedule create "every Monday at 9am" "calmhive afk 'weekly dependency audit and update' --iterations 8"
calmhive schedule list # See all scheduled tasks

Control with your voice

calmhive voice # Say "hey friend, analyze this codebase"

Monitor everything

calmhive tui # See all running sessions in one place
`

🚀 Pro User Shortcuts

Morning workflow - Check overnight AFk sessions:
`bash
calmhive t # Visual overview + navigate sessions with j/k
calmhive c -c # Continue last conversation to see what happened
`

Power user aliases (all tested and working):
`bash
calmhive c "debug this error" # Chat alias
calmhive a "refactor codebase" # AFk alias
calmhive r "write unit tests" # Run alias
calmhive t # TUI alias
calmhive v # Voice alias
calmhive s list # Schedule alias (NEW)
calmhive tmp list # Template alias (NEW)
calmhive p status # Progress alias (NEW)
`

AFk session management:
`bash
calmhive a status -d # Detailed status with progress info
calmhive a tail abc12345 # Live logs from specific session
calmhive a stop abc12345 # Stop specific session
calmhive a cleanup 3 # Clean up sessions older than 3 days
calmhive a killorphans # Clean up any stuck processes
`

Advanced patterns:
`bash


Long-running overnight tasks

calmhive a "implement OAuth 2.0 flow" --iterations 20 --model sonnet

Continue from where AFk left off

calmhive c -c "What did you accomplish? Show me the changes."

Quick pipeline: develop → test → review

calmhive r "add feature X"
calmhive a "write comprehensive tests" --iterations 5
calmhive c "review the implementation and suggest improvements"
`

📚 Commands

$3

Talk to Claude with full access to 86 tools (15 core + 71 MCP integrations). Supports all Claude flags.
`bash
calmhive chat # Start interactive session
calmhive c "debug this error" # Quick question
calmhive chat -c # Continue previous conversation
`

💡 Pro Tip: Plan Mode (Shift+Tab)
In interactive chat sessions, press Shift+Tab to enter plan mode - Claude will research and create a detailed plan before executing complex tasks.

$3

Wrapper for claude run with automatic tool approval. Perfect for scripts and CI/CD.
`bash
calmhive run "add type annotations to services/auth.js"
calmhive r "find and fix n+1 queries in the API"
calmhive run "migrate database schema to TypeScript"
`

$3

Run multiple Claude sessions sequentially with automatic retry on usage limits.
`bash
calmhive afk "refactor entire codebase" --iterations 20 # Auto sleep prevention
calmhive afk "overnight task" --iterations 50 # Max: 69 iterations
calmhive afk "quick fix" --iterations 3 --no-prevent-sleep # No caffeinate
calmhive afk status # Check all running tasks
calmhive afk status -d # Detailed view with full task info
calmhive afk tail abc-123 # Watch live progress
calmhive afk stop abc-123 # Stop a task
calmhive afk cleanup # Clean old sessions (7 days)
calmhive afk cleanup --dry-run # Preview cleanup safely
calmhive afk cleanup 30 --force # Custom retention, skip prompts
`

$3

Full-featured voice interface with speech recognition and text-to-speech.
`bash
calmhive voice # Start voice interface
calmhive voice --debug # Debug mode

Say "hey friend", "calmhive", or "ok friend" to activate

`

$3

Revolutionary AI performance through beautiful parallel execution patterns. No more watching AI work inefficiently one task at a time!

What You'll See:
When AI analyzes your codebase, instead of slow sequential processing, you'll see elegant parallel cascades:
`
⏺ Task(Analyze authentication patterns) ⎿ Initializing…
⏺ Task(Update security documentation) ⎿ Initializing…
⏺ Task(Fix validation errors) ⎿ Initializing…
⏺ Task(Add missing test coverage) ⎿ Initializing…
`

Performance Impact:
- 4-10x Speed Improvement: Multiple analysis tasks run simultaneously
- Visual Beauty: Watch AI work optimally with cascade displays
- Trust Building: See AI efficiency in action, not sequential plodding
- Self-Correction: AI detects and prevents inefficient patterns

How It Works:
The v14.4.0 CLAUDE.md enhancement automatically:
- Detects when multiple tasks can run in parallel
- Blocks inefficient sequential processing
- Creates beautiful visual feedback for parallel execution
- Provides self-correction when violations are attempted

User Benefits:
- Faster Results: Complex analysis completes in 1/4 the time
- Better Experience: Watch AI work efficiently, not waste time
- Automatic Optimization: No configuration needed--it just works
- Transparent Performance: See exactly how AI is optimizing your workflow

Example Workflow Improvement:
`bash


Complex codebase analysis now runs optimally:

calmhive chat "analyze the security, performance, and testing coverage of this codebase"

Before v14.4.0: Sequential plodding (12+ minutes)

After v14.4.0: Beautiful parallel execution (3-4 minutes)

`

This enhancement works across all Calmhive commands--chat, run, afk, and voice--ensuring you always get optimal AI performance.

$3

Schedule recurring tasks using natural language powered by Claude Code.
`bash
calmhive schedule create "daily at 9am" "calmhive chat 'morning standup'"
calmhive schedule create "every Monday at 2pm" "calmhive afk 'weekly report' -i 5"
calmhive schedule list # View all schedules
calmhive schedule restore # Start scheduler service
calmhive schedule update --time "every 6 hours" # Update schedule
calmhive schedule stop # Stop a specific schedule
calmhive schedule delete # Remove a schedule
`

$3

Pre-built workflows with variable substitution for common development tasks.

Core Template System:
`bash
calmhive template list # See available templates
calmhive template show bug-fix # View template details with variables
calmhive template apply bug-fix ISSUE_DESCRIPTION="memory leak in auth service"
`

Available Templates:
- bug-fix: Systematic bug investigation (8 iterations)
- feature-development: End-to-end implementation (12 iterations)
- refactoring: Safe code refactoring (10 iterations)
- testing-setup: Test infrastructure (6 iterations)
- performance-optimization: Performance tuning (8 iterations)

Template Integration with AFk:
`bash


Use templates with AFk sessions for complex automation

calmhive afk "$(calmhive template apply performance-optimization TARGET_SYSTEM=api-server)" --iterations 10

Schedule template-driven workflows

calmhive schedule create "weekly on Friday at 4pm" \
"$(calmhive template apply bug-fix ISSUE_DESCRIPTION='weekly maintenance check')"
`

Variable Substitution:
Templates support dynamic variables for customization:
- ISSUE_DESCRIPTION - Describe the specific problem
- TARGET_SYSTEM - Specify which system/component to work on
- PRIORITY_LEVEL - Set urgency (high/medium/low)
- ITERATIONS - Override default iteration count

$3


Monitor detailed progress of AFk sessions with iteration summaries.
`bash
calmhive progress # View iteration summaries and milestones

Shows:

- Iteration-by-iteration summaries

- Milestones achieved

- Actions taken per iteration

- Overall progress status

`

$3

Automatically update to the latest version from npm.
`bash
calmhive update # Update to latest version
`

$3

Beautiful interface to monitor and manage all sessions with smart log resolution.
`bash
calmhive tui # Launch the UI

Use arrow keys to navigate, Enter to view logs

Now automatically finds logs from V3, AFk, or legacy locations

`

🖥️ TUI Features:
- Real-time Monitoring - Watch sessions update live
- Vim-style Navigation - j/k to move, gg/G for top/bottom
- Session Control - Stop (s), Resume (r), or view logs (Enter)
- Smart Search - Press / to search through sessions
- Progress Tracking - Visual progress bars and ETAs
- Log Viewer - Scroll through detailed session logs

Keyboard Shortcuts:
- j/k or ↑/↓ - Navigate up/down
- Enter - View session logs
- s - Stop selected session
- r - Resume paused session
- gg - Jump to first session
- G - Jump to last session
- / - Search sessions
- q or Esc - Exit TUI

🔥 Killer Features

$3

Ever had an overnight task die at iteration 10 due to usage limits? Not anymore!

`
Iteration 10 ✓
⚠️ Usage limit detected
⏳ Waiting 60s before retry...
⏳ Waiting 120s before retry...
✅ Iteration 11 started!
`

Calmhive automatically detects and handles usage limits with exponential backoff.

$3

Your MacBook won't sleep during long AFk sessions:

`
☕ Sleep prevention enabled (caffeinate PID: 12345)
🚀 Starting iteration 1 of 50
... runs all night without interruption ...
☕ Sleep prevention disabled (caffeinate stopped)
`

Automatic for sessions >5 iterations, disable with --no-prevent-sleep.

$3

When Claude hits context limits, Calmhive automatically attempts multiple strategies to compress the conversation:

`
⚠️ Context limit approaching
🗜️ Attempting context compression...
✅ Context compressed successfully!
`

Features:
- Multiple /compact command formats for compatibility
- Automatic retry with different compression strategies
- Context usage monitoring and reporting
- Fallback manual compression if needed

🚦 What's Next?

After installation:
1. Run calmhive chat to start interactive mode
2. Try calmhive run "create a hello world script" for automation
3. Use calmhive voice to enable voice control
4. Explore with calmhive --help

🛠️ Advanced Usage

$3


`bash
calmhive chat --model opus "complex analysis"
calmhive run --model sonnet "standard task"
`

$3

`bash
cat code.js | calmhive chat -p "review this"
calmhive run "generate SQL schema" > schema.sql
`

$3

`bash

Simple prompt mode

calmhive afk "task" --simple --iterations 5

Advanced mode with sequential thinking

calmhive afk "complex task" --custom-steps --iterations 25
`

🔧 Installation Troubleshooting

Command not found after npm install?
`bash


Add npm global binaries to PATH:

export PATH="$PATH:$(npm prefix -g)/bin"

Or find where npm installs global packages:

npm prefix -g
`

Permission errors with npm?
`bash


Use npx to run without global install:

npx @calmhive/calmhive-cli chat "hello"

Or use the git clone method above for full control

`

Git clone setup example:
`bash


Create repos directory if it doesn't exist

mkdir -p $HOME/repos

Clone and setup

git clone https://github.com/joryeugene/calmhive-cli.git $HOME/repos/calmhive-cli

Add to PATH (macOS/zsh)

echo 'export PATH="$HOME/repos/calmhive-cli/bin:$PATH"' >> ~/.zshrc
source ~/.zshrc

For bash users (or Linux):

echo 'export PATH="$HOME/repos/calmhive-cli/bin:$PATH"' >> ~/.bashrc

source ~/.bashrc

Test it works

calmhive --help
`

Claude CLI missing?
Install from claude.ai/claude-cli

Best with Claude Max!
While Pro/Teams work great, Claude Max offers the highest rate limits for heavy automation workflows.

Voice not working?
`bash


Install uv package manager if needed (macOS):

curl -LsSf https://astral.sh/uv/install.sh | sh

Set OpenAI API key:

export OPENAI_API_KEY=your_key_here

Start voice interface:

calmhive voice
`
> 📱 Platform Note: Voice features are optimized for macOS. Linux/Windows compatibility may vary.

MCP Tools not available?
Calmhive requires Claude Max, Pro, or Teams subscription for MCP (Model Context Protocol) tools. With Claude Free, basic chat/run commands work but advanced features are limited.

Need detailed MCP setup?
See our comprehensive MCP Setup Guide for step-by-step instructions on configuring essential tools like Sequential Thinking and OmniSearch.

💡 Tips & Tricks

$3

Intelligent Context Preservation Without Spam

Calmhive v14.2.1 introduces a revolutionary smart injection system that maintains CLAUDE.md context while eliminating console spam:

#### 🔍 How Smart Injection Works
`bash


Enable debug mode to see the magic happen


CALMHIVE_DEBUG=1 calmhive chat "hello world"
`

Request Analysis in Action:
`
[Calmhive Debug] === REQUEST ANALYSIS ===
[Calmhive Debug] Request has tools: false
[Calmhive Debug] Messages count: 1
[Calmhive Debug] Has recent tool use: false
[Calmhive Debug] Request analysis: {
"requestType": "fresh-user-message",
"isUserMessage": true
}
[Calmhive Debug] Injection decision: INJECT (fresh-user-message)
[Calmhive Interceptor] Rules injected (injection #1)
`

#### ⚡ Smart Features
- Request Type Detection: Distinguishes user messages from tool calls
- Message Deduplication: Same message content = single injection
- Tool Context Awareness: Skips injection during tool execution

#### 📊 Debug Mode Features
Enable with CALMHIVE_DEBUG=1 to see:
- Full request body analysis
- Injection decision reasoning
- Message deduplication in action
- Request type classification

$3

Create a ~/.claude/CLAUDE.md file to provide persistent context across all Claude sessions:
- Global coding preferences and standards
- Frequently used tools and workflows
- Custom instructions for your AI assistant

The enhanced CLAUDE.md includes hierarchical decision framework to prevent catastrophic mistakes.

Quick Setup:
`bash


Install the optimized CLAUDE.md template

calmhive config install

Test smart injection

calmhive chat "analyze this codebase"
`

Learn more in our Smart Injection Documentation.

$3

Calmhive v3 includes curated command templates for common development workflows:

Core Templates Available:
- bug-hunting.md - Systematic bug detection and resolution
- documentation.md - Technical documentation generation
- refactoring.md - Code improvement and architectural changes
- test-generation.md - Test suite creation with coverage analysis

Expert Engineering Personas:
- expert-frontend-engineer.md - Modern web development expertise
- expert-backend-engineer.md - Server architecture and APIs
- expert-devops.md - Infrastructure automation and deployment

Using Templates with Calmhive:
`bash


Templates are .md files that need to be copied to Claude's commands directory

cp -r "$(npm root -g)/@calmhive/calmhive-cli/commands/"*.md ~/.claude/commands/

Then use them in interactive chat sessions with /user:

calmhive chat
> /user:bug-hunting
> Find memory leaks in this application

Or load expert personas:

> /user:expert-frontend-engineer
> Optimize this React component for performance
`

Note: Templates can only be invoked with /user:command-name in interactive sessions.

$3

Sleep Prevention
`bash


Automatic for >5 iterations - your Mac stays awake!

calmhive afk "big project" --iterations 20 # ☕ caffeinate enabled

Disable if you prefer manual sleep control

calmhive afk "small task" --iterations 8 --no-prevent-sleep
`

Stopping AFk Sessions
`bash


❌ DON'T use Ctrl+C (won't work - AFk runs in background)

✅ DO use proper stop command:

calmhive afk stop
`

Continue Where You Left Off
`bash


After an AFk session completes, pick up the conversation:


calmhive chat -c # Claude remembers the context!
`

Monitor Progress
`bash
calmhive afk status # See all sessions
calmhive afk tail # Watch live logs
calmhive tui # Visual interface
`

$3

Mixed Console Output
If you see iterations jumping (1 → 3), you may have multiple AFk processes:
`bash
ps aux | grep "commands/afk" # Check for duplicates
kill # Kill orphan processes
`

Long Tasks Timing Out
AFk sessions can run for hours - no artificial timeouts!
`bash


These will run as long as needed:


calmhive afk "complex refactoring" -i 30
calmhive afk "comprehensive testing" -i 69 # Max iterations increased!
`

Usage Limit Recovery
Calmhive automatically handles rate limits with exponential backoff:
`
✅ Iteration 5 complete
⚠️ Usage limit detected
⏳ Waiting 30s → 1m → 2m → 4m...
✅ Iteration 6 started!
`

$3

Real Morning Workflow (verified commands)
`bash


1. Check what happened overnight

calmhive t # TUI shows all sessions, navigate with j/k, Enter for details

2. Continue conversation from where AFk left off

calmhive c -c "What did you accomplish? Show me the changes."

3. View logs from specific session (if needed)

calmhive a tail afk-abc12345 # Live/historical logs

4. Clean up old sessions periodically

calmhive a cleanup 7 # Remove sessions older than 7 days
`

Production-Ready Automation Workflows
`bash


Daily ESLint Deep Scan & Auto-Fix (comprehensive code quality)

calmhive schedule create "every day at 6am" \
"calmhive afk 'Perform comprehensive ESLint analysis on the entire codebase. Run eslint on all JavaScript files in lib/, cmd/, and test/ directories to get a full diagnostic report. Categorize all warnings by type (no-unused-vars, prefer-const, no-case-declarations, etc). For the top 5 most common warning types, systematically fix 10 instances of each type, ensuring proper variable usage or safe removal of dead code. After each fix, run the specific test file related to that module to ensure no regressions. Create a detailed report showing: total warnings before/after, warnings fixed by category, any code patterns that need manual review. If any tests fail, revert that specific fix and document why it needs manual intervention. Finally, commit all successful fixes with a message detailing each category of fixes made.' --iterations 10"

Weekly Security Audit & Patching

calmhive schedule create "every Monday at 8am" \
"calmhive afk 'Conduct comprehensive security audit of all npm dependencies. Run npm audit to identify vulnerabilities. For each vulnerability found, research the security advisory details, understand the attack vector, and evaluate the impact on our codebase. For critical and high severity issues, check if patches are available and test them in isolation. Update all dependencies that have security patches available, running the full test suite after each update to ensure compatibility. Generate a security report including: vulnerabilities found, patches applied, dependencies that need manual updates, and any temporary mitigations implemented. Commit the updated package-lock.json with detailed notes about each security fix.' --iterations 8"

Template-driven bug investigation

calmhive afk "$(calmhive template apply bug-fix \
ISSUE_DESCRIPTION='memory leak in auth service')" --iterations 8

Schedule template-based workflows

calmhive schedule create "every Friday at 4pm" \
"calmhive afk '$(calmhive template apply performance-optimization \
TARGET_SYSTEM=api-server)' --iterations 8"
`

TUI Power Navigation (tested shortcuts)
`bash
calmhive t # Launch TUI


Inside TUI:

gg = Go to top, G = Go to bottom

j/k = Navigate up/down (vim style)

Enter = View session details

l = View logs, s = Stop session

q = Quit, n = New session

`

Smart Model Selection
`bash
calmhive chat --model sonnet "standard work" # Balanced (default)
calmhive afk --model sonnet "standard work" -i 5 # Balanced performance
calmhive chat --model opus "complex analysis" # Most capable
`

Pipe Workflows
`bash


Process files through Claude

find . -name "*.js" | calmhive chat -p "list potential bugs in these files"

Generate and save output

calmhive run "create comprehensive test suite" > test-plan.md
`

Using Prompt Files
`bash


Pass a markdown file as AFk task description

calmhive afk "$(cat prompt.md)" --iterations 20

Use complex prompts from files

calmhive chat "$(cat detailed-instructions.md)"
calmhive run "$(cat build-script-prompt.md)"
`

Voice + AFk Combo
`bash


Start voice interface

calmhive voice

Say: "Start an AFk session to refactor the entire authentication system with 20 iterations"

Then monitor with TUI while doing other work!

`

$3

- Context Continuity: Use chat -c after AFk to maintain conversation history
- Background Monitoring: Keep calmhive tui open in a separate terminal
- Smart Iterations: 5-15 iterations for most tasks, 20+ for major refactoring
- Clean Shutdown: Always use afk stop instead of killing terminals
- Model Efficiency: Sonnet for most work (default), Opus for complex reasoning

🎯 Power User Configuration

Calmhive includes an advanced CLAUDE.md.example file with powerful directives that transform Claude's behavior:

$3

- Adaptive Retry Logic: Automatically recovers from Claude usage limits with exponential backoff
- Background Processing: AFk mode runs long tasks while you work on other things
- Voice Activation: Say "hey friend" to control Claude hands-free
- Process Monitoring: Beautiful TUI to track all your automation sessions
- MCP Integration: 70+ tools including Sequential Thinking, GitHub, Playwright, and more

$3

`bash

1. Install the enhanced CLAUDE.md configuration

calmhive config install

2. Configure available tools based on your MCP setup

Check what MCP tools you have configured

claude mcp list

3. Test the enhanced behavior

calmhive chat "analyze this codebase" # Watch the difference!
`

The CLAUDE.md configuration transforms Claude into a reliable automation partner that:
- ✅ Searches before making destructive changes
- ✅ Tests solutions before claiming success
- ✅ Uses configuration over deletion for cleanup tasks
- ✅ Shows actual evidence instead of assumptions
- ✅ Recovers gracefully from failures
- ✅ Integrates seamlessly with MCP tools when available

🔗 Ecosystem Integration

$3

- Calmhive.com - Main website with full documentation, demos, and getting started guides
- Prompthive.sh - Perfect companion tool for prompt management and complex automation workflows

$3

Prompthive.sh is the ideal companion to Calmhive, providing sophisticated prompt management, workflow templates, and automation patterns. Together they create a complete AI development platform.

What Prompthive.sh Adds:
- 🎯 Advanced prompt engineering tools
- 📚 Workflow template libraries
- 🔄 Complex automation patterns
- 🧠 Prompt optimization frameworks

Integration Examples:
`bash


Combined workflow: Prompt management + AI automation

prompthive.sh execute security-audit.yaml && \
calmhive afk "analyze and fix security issues found" --iterations 8

Schedule integrated workflows

calmhive schedule create "daily at 6am" \
"prompthive.sh execute daily-review.yaml && calmhive run 'summarize findings > reports/daily.md'"

Template-driven with prompt management

calmhive schedule create "weekly on Friday at 4pm" \
"$(calmhive template apply performance-optimization TARGET_SYSTEM=api-server) && prompthive.sh execute performance-report.yaml"
`

Why Use Both Together:
- Prompthive.sh: Manages complex prompts, templates, and workflows
- Calmhive: Provides AI automation, scheduling, and background processing
- Combined: Complete automation ecosystem for development workflows

🐝 AI Guidelines for Teams

$3

Calmhive includes powerful AI guidelines that transform how AI assists your team. These guidelines work with Claude Code CLI, ChatGPT, Claude.ai, Cursor, and other AI interfaces.

What's New in v15.1.0:
- ✅ Root Cause Analysis Integration - Forces 5 Whys methodology before jumping to solutions
- ✅ 5 Key Engineering Lessons - Prevents over-engineering and maintains focus
- ✅ Package-First Development - Mandatory package search before custom implementations
- ✅ Voice Authenticity - Eliminates corporate speak and AI giveaways
- ✅ Evidence-Based Responses - No claims without proof

$3

We provide two versions optimized for different use cases:

CLI Version (CLAUDE.md) - For Claude Code CLI users:
- Full Task parallelization support
- TodoWrite/Read integration
- Command-line specific optimizations
- MCP tool awareness
- Located: ~/.claude/CLAUDE.md

Desktop Version (CLAUDE-DESKTOP.md) - For web interfaces:
- Works with ChatGPT, Claude.ai, Cursor, etc.
- No CLI-specific features
- Same behavioral improvements (RCA, Key Lessons, etc.)
- Customizable for team culture
- Example: /calmhive/v3/config/CLAUDE-DESKTOP.md.example

$3

While the guidelines work great out-of-the-box, teams can adapt them:

1. Voice Traits - Adjust formality level to match team culture
2. Evidence Standards - Define what constitutes "proof" in your context
3. Key Lessons - Add team-specific lessons learned
4. Technical Standards - Include your coding conventions

The guidelines are designed to enhance, not replace, your team's existing practices.

📖 Documentation

- Calmhive.com - Official website with demos, features showcase, and getting started guides
- Prompthive.sh - Companion tool for advanced prompt management and workflow automation
- Calmhive Blog Post - Full story behind Calmhive
- Terminal Velocity Guide - CLI-first AI development workflow
- Architecture Guide - Detailed command documentation
- MCP Setup Guide - Configure powerful MCP tools
- Scheduling Examples - Production-ready automation workflows
- README Templates - Template file documentation
- GitHub Repository - Source code & issues
- npm Package - Version history

🛡️ Security & Responsible Usage

⚠️ IMPORTANT: Calmhive gives Claude direct access to your system through powerful automation tools. While this enables incredible productivity, it also means Claude can execute any command, modify files, or make network requests. A misunderstood instruction could potentially cause data loss or system changes.

✅ Essential safety practices:
- Backup your data - Maintain regular backups of important files and projects
- Use version control - Keep your code in Git repositories with remote backups
- Review automation - Monitor what commands Claude executes, especially in AFk mode
- Start small - Test with simple tasks before running complex automations
- Monitor tool usage - Some MCP tools have their own API costs (search engines, etc.)

🔒 Built-in safeguards:
- All commands are logged for review (calmhive tui to monitor)
- Background tasks can be stopped anytime (calmhive afk stop `)
- No elevated permissions required (runs as your user)
- Process isolation for AFk sessions

> 💡 Bottom line: Calmhive is a powerful automation tool that amplifies both your capabilities and potential risks. Use it thoughtfully, maintain good backup habits, and start with simple tasks to build confidence.

💻 Platform Compatibility

Primary Platform: macOS
- ✅ Fully tested and optimized
- ✅ All features supported (voice, TUI, AFk background processing)
- ✅ Native integrations (caffeinate for sleep prevention, etc.)

Linux: Limited support
- ⚠️ Basic commands work (chat, run, afk)
- ⚠️ Voice features may require additional setup
- ⚠️ TUI navigation tested but some features may vary
- ⚠️ Background processing should work but less tested

Windows: Experimental
- ⚠️ Core functionality may work with WSL
- ❌ Voice features not supported
- ❌ Some shell-specific features may not work
- ❌ Not recommended for production use

> Recommendation: For the best Calmhive experience, use macOS. Linux users can try it but expect some rough edges. Windows users should consider WSL or wait for better cross-platform support.

📚 Documentation

- Architecture Guide - Technical overview and command structure
- MCP Setup Guide - Complete MCP tools configuration
- README Templates - Template file documentation
- Claude Code Best Practices - Comprehensive Claude Code usage guide

🤝 Contributing

PRs welcome!

📄 License

MIT © 2025 Jory Pestorious

---

lets bee friends 🐝

> The intelligent Claude CLI that works when Claude doesn't