v4.12.0 - Specialist Agents & Claude Code Parallel Fix

🎉 New Features

6 New Specialist Agents

Now available for both Claude Code and Droid CLI:

Specialist	Purpose
backend-specialist	APIs, authentication, server-side logic, microservices
frontend-specialist	React components, state management, styling, accessibility
database-specialist	Schema design, migrations, query optimization, SQL/NoSQL
devops-specialist	CI/CD pipelines, Docker, cloud infrastructure, monitoring
test-specialist	Unit tests, integration tests, E2E tests, test strategies
full-stack-specialist	End-to-end feature implementation across all layers

Use with /orchestrate-tasks to assign specialists to task groups for parallel execution.

---

🔧 Critical Fix: Claude Code Parallel Execution

Before (Broken)

Claude Code /implement-tasks was incorrectly using droid exec (Factory CLI method).

After (Fixed)

• Uses native Claude Code Task tool for parallel execution
• Spawns assigned specialists from orchestration.yml
• Falls back to implementer agent when specialists don't exist
• Removed linear dependency blocker - proceeds with parallel execution

---

Platform Behavior

Platform	Parallel Method	Agents
Claude Code	Task tool → spawn specialists	6 new + existing
Droid CLI	droid exec shell script	6 new + existing

Droid CLI Specialists Include:

• ✅ Color field for CLI display
• ✅ TodoWrite progress tracking
• ✅ No tool restrictions (inherits all tools)

---

Usage

# Assign specialists to task groups
/orchestrate-tasks

# Run implementation with parallel execution (Option A)
/implement-tasks

The specialists will be automatically invoked based on orchestration.yml assignments.

v4.11.1 - Claude Code Native Parallel Execution Fix

🔧 Critical Fix: Claude Code Parallel Execution

Fixed

• CRITICAL: Claude Code /implement-tasks parallel execution now uses correct method
  • Was incorrectly using droid exec (Factory CLI method) in Claude Code
  • Now uses native Claude Code Task tool with multiple parallel subagent invocations
  • Spawns the assigned specialists from orchestration.yml, not generic agents

Changed

• Claude Code implement-tasks.md - Option A (Parallel Execution) completely rewritten:

  • Requires /orchestrate-tasks to be run first to assign specialists
  • Reads orchestration.yml to get assigned_specialist for each task group
  • Verifies specialists exist in .claude/agents/
  • Spawns all specialists in a SINGLE message for true parallelism
  • Includes proper standards resolution from orchestration.yml
  • Clear error messages if orchestration.yml or specialists are missing

• Droid CLI implement-tasks.md - Unchanged (still uses droid exec with Factory API)

Platform Behavior Summary

Platform	Parallel Method	Prerequisite
Claude Code	Task tool → spawn specialists	/orchestrate-tasks to assign specialists
Droid CLI	droid exec shell script	Factory API key

Impact

• ✅ Claude Code users get proper native parallel execution
• ✅ Factory/Droid CLI users unaffected (still use droid exec)
• ✅ Specialists from /orchestrate-tasks are properly utilized
• ✅ No more confusion about which parallel method to use per platform

v4.11.0 - Standards Compliance & Package Manager Detection

What's New

Standards Compliance

All workflow commands (/shape-spec, /write-spec, /create-tasks, /orchestrate-tasks) now properly read and adhere to project standards:

• Reads ALL standards recursively from droidz/standards/**/*.md
• Includes any custom standards directories you create
• Standards are applied at every step of the workflow

Package Manager Detection

Commands now automatically detect your project's package manager and use the correct commands:

Lockfile	Package Manager
bun.lockb	bun
pnpm-lock.yaml	pnpm
yarn.lock	yarn
package-lock.json	npm

This affects:

• /create-tasks - Generated tasks use correct package manager commands
• /orchestrate-tasks - Implementation prompts use correct commands
• /implement-tasks - All execution uses correct package manager
• Implementation verification - Tests run with correct package manager
• CI/CD pipeline generation - Workflows use correct commands

Files Updated

• 18 command/agent files across claude and droid_cli payloads
• Installer version bumped to 4.11.0

Installation

bash <(curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh)

Or update existing installation by re-running the installer.

v4.9.0 - Comprehensive Documentation & Platform Fixes

📚 v4.9.0 - Comprehensive Documentation & Platform-Agnostic Updates

Major Documentation Overhaul

Complete README rewrite: 554 → 1,219 lines of comprehensive documentation!

This release makes Droidz significantly more accessible with detailed guides, examples, and best practices.

---

🆕 What's New

1. Complete Commands Reference

Every command now documented in detail:

📋 /standards-shaper ⭐ (Start Here)

• Purpose: Create project-wide coding standards
• When to use: Once per project, before development
• Creates: standards/ with global, backend, frontend, testing

📋 /plan-product

• Purpose: Define product vision and roadmap
• Creates: mission.md, roadmap.md, tech-stack.md

📋 /shape-spec

• Purpose: Shape requirements for new features
• Creates: planning/requirements.md, decisions.md

📋 /write-spec

• Purpose: Write detailed specification
• Creates: Complete spec.md with architecture, APIs, testing

📋 /create-tasks

• Purpose: Break spec into implementable tasks
• Creates: tasks.md with task groups and dependencies

📋 /orchestrate-tasks

• Purpose: Plan multi-agent implementation
• Creates: orchestration.yml, implementation prompts

📋 /implement-tasks

• Purpose: Execute implementation
• Modes: Automated (subagents) or Manual (prompts)

📋 /improve-skills

• Purpose: Enhance AI agent skills

Each command includes:

• ✅ Purpose and when to use
• ✅ Step-by-step breakdown
• ✅ Output files created
• ✅ Real conversation examples

---

2. All 8 Agents Documented

Complete profiles for:

• product-planner - Product strategy and roadmaps
• spec-shaper - Requirements gathering and research
• spec-writer - Detailed specification writing
• spec-verifier - Specification review and validation
• spec-initializer - Quick spec folder setup
• tasks-list-creator - Breaking specs into tasks
• implementer - Feature implementation
• implementation-verifier - Testing and verification

Each agent includes:

• Expertise area
• When to use
• What it creates
• Best use cases

---

3. Best Practices Guide

✅ 8 DOs with Rationale:

1. Start with standards (/standards-shaper first!)
2. Always shape before writing specs
3. Break down complex tasks into small pieces
4. Use orchestration for big features
5. Reference specs in code comments
6. Update standards as you learn
7. Keep specs updated during implementation
8. Use task groups for organization

❌ 8 DON'Ts with Corrections:

1. Don't skip planning → Use the workflow
2. Don't write vague specs → Be specific
3. Don't ignore standards → Follow patterns
4. Don't mix concerns → Separate task groups
5. Don't lose context → Reference specs
6. Don't skip testing → Tests = implementation
7. Don't work in isolation → Use workflow for collaboration
8. Don't forget to commit → droidz/ in git!

---

4. Real-World Examples

Example 1: Building Real-Time Notifications
Complete walkthrough from standards to shipping:

> /standards-shaper
> /plan-product
> /shape-spec
> /write-spec
> /create-tasks
> /orchestrate-tasks
# → Ship! 🚀

Example 2: Onboarding New Team Member
Day 1 productivity guide:

• Clone repo
• Install Droidz
• Read product docs
• Review standards
• Start first feature

Example 3: Mid-Project Standards
Adding standards to existing codebase:

> /standards-shaper
# Detects: React 18, TypeScript, Tailwind, tRPC, Prisma
# Creates standards matching YOUR code style

---

5. Enhanced Documentation

New Sections:

• 📖 Table of Contents for easy navigation
• 🎯 "Why Droidz?" - Problem/Solution breakdown
• 🔄 Visual workflow diagram
• 🛠️ Enhanced troubleshooting guide
• 🎨 Better formatting and organization

Preserved Sections:

• ✅ Discord community links
• ✅ PayPal donation information
• ✅ Supporters acknowledgment
• ✅ Platform support table

---

🔧 Platform-Agnostic Fixes

orchestrate-tasks Command

Fixed: Platform-specific terminology causing confusion

Before:

task_groups:
  - name: authentication
    claude_code_subagent: backend-specialist  # ❌ Claude-specific

After:

task_groups:
  - name: authentication
    assigned_specialist: backend-specialist  # ✅ Works everywhere

Changes:

• "Claude Code subagent" → "specialist"
• YAML key claude_code_subagent → assigned_specialist

Now works correctly for:

• ✅ Factory AI (droids)
• ✅ Claude Code (subagents)
• ✅ Cursor, Cline, and other platforms

---

📊 Documentation Comparison

Metric	Before	After	Change
Total Lines	554	1,219	+120%
Commands Documented	Basic list	8 detailed guides	Full coverage
Agent Profiles	Brief list	8 complete profiles	Comprehensive
Examples	1 basic	3 real-world	Practical
Best Practices	None	16 (8 DOs + 8 DON'Ts)	Added
Troubleshooting	4 items	6 detailed solutions	Enhanced

---

🚀 Getting Started

New Users

# 1. Install
cd your-project
bash <(curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh)

# 2. Start with standards
> /standards-shaper

# 3. Follow the workflow
> /plan-product
> /shape-spec
> /write-spec
# ... continue through phases

Existing Users

New resources available:

• Complete command reference in README
• Best practices guide
• Real-world examples to follow
• Enhanced troubleshooting

---

📦 Installation

# Fresh install
bash <(curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh)

# Upgrade existing
cd your-project
bash <(curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh)

---

🎯 Impact

For New Users:

• ✅ Clear understanding of what Droidz does
• ✅ Step-by-step command guides
• ✅ Know when to use each tool
• ✅ Real examples to follow

For Existing Users:

• ✅ Complete command reference
• ✅ Best practices to implement
• ✅ Troubleshooting solutions
• ✅ Customization examples

For Contributors:

• ✅ Better onboarding documentation
• ✅ Clear framework structure
• ✅ Examples to build upon

---

Full Changelog: https://github.com/korallis/Droidz/blob/main/CHANGELOG.md

This release represents a major step forward in making Droidz accessible and powerful for all users! 📚✨

v4.8.1 - Critical Fix: All Droids Now Functional

🔧 v4.8.1 - Critical Bug Fix: Droids Now Work Without Failures

Critical Fix

All 8 custom droids were failing due to missing workflow files. This release fixes all droids to work immediately after installation.

Problem

Users reported: "droids reference things that don't exist" causing failures like:

• task-list-creator droid failed
• spec-writer droid failed
• All custom droids showed "file not found" errors

Root Cause

All droids used {{workflows/...}} template syntax to reference external workflow files that never existed. The installer copied files as-is without processing templates, leaving broken file references.

Solution

✅ Inlined complete workflow instructions directly into each droid
✅ Removed all broken template references (38 files fixed)
✅ Droids now work immediately after installation

Fixed Droids (All 8)

1. tasks-list-creator

• Added complete task breakdown process (6 steps)
• How to analyze specs and create actionable tasks

2. spec-writer

• Added specification writing guidelines (8 sections)
• Complete template for writing specs

3. product-planner

• Added product planning phases (4 phases)
• From gathering info to defining tech stack

4. implementer

• Added implementation workflow (7 steps)
• From understanding task to verifying completion

5. spec-shaper

• Added specification shaping process (6 steps)
• Research, clarify, define, structure

6. spec-initializer

• Added spec initialization steps (4 steps)
• Gather info, create structure, identify stakeholders

7. spec-verifier

• Added verification checklist (5 checks)
• Completeness, clarity, implementability

8. implementation-verifier

• Added verification process (4 steps)
• Verify tasks, update roadmap, run tests, create report

Files Fixed

• 16 droid files (8 droids × 2 platforms)
• 22 command files (11 commands × 2 platforms)
• Total: 38 files with broken references resolved

Verification

✅ 0 broken template references remaining
✅ All 8 droids functional
✅ All 11 commands working
✅ Both platforms synced (Factory Droid CLI + Claude Code)

Installation

Fresh Install

curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh | bash

Upgrade

cd your-project
curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh | bash

What Changed

Template References Removed:

• {{workflows/planning/*}} - Product planning workflows
• {{workflows/specification/*}} - Specification workflows
• {{workflows/implementation/*}} - Implementation workflows
• {{standards/*}} - Standards (now using skills instead)
• {{UNLESS}}/{{IF}} - Unprocessed conditional blocks

Inlined Content:

• Complete step-by-step workflows in each droid
• Self-contained instructions
• No external dependencies

Impact

Before: ❌ All custom droids failed with "file not found"
After: ✅ All custom droids work immediately after installation

---

Full Changelog: https://github.com/korallis/Droidz/blob/main/CHANGELOG.md

v4.8.0 - All 50 Skills Comprehensively Optimized 🎉

🎉 v4.8.0 - All Skills Professionally Optimized to Maximum Quality

🌟 Major Quality Enhancement

Every single skill in Droidz has been comprehensively optimized with detailed descriptions and extensive usage examples!

This release represents a complete transformation of the skills library, bringing professional quality and maximum discoverability to all 50 skills.

✨ What's New

📚 Complete Skills Library Optimization

Every skill now features:

• Comprehensive Descriptions: 200-400+ character detailed, activation-oriented descriptions
• Extensive Examples: 12+ specific "When to use this skill" examples per skill
• Consistent Structure: All skills follow skill-creator patterns throughout
• Maximum Clarity: No length constraints - descriptions are as thorough as needed
• Professional Quality: Production-ready documentation across entire library

🔥 Quality Comparison

Before (Generic, 3 examples):

description: E2E testing, browser automation, web scraping.

## When to use this skill
- E2E testing, browser automation, web scraping.
- When working on related tasks or features
- During development that requires this expertise

After (Comprehensive, 13+ examples):

description: Automate browser testing, web scraping, and user workflow testing 
  with Playwright across multiple browsers. Use when writing end-to-end tests 
  for web applications, automating repetitive browser tasks, scraping data from 
  websites, testing across Chrome/Firefox/Safari, taking screenshots for visual 
  regression testing, testing authentication flows, filling and submitting forms 
  programmatically, testing responsive designs across viewports, or any browser 
  automation requiring reliable, cross-browser testing capabilities.

## When to use this skill
- Writing end-to-end tests for web applications
- Automating repetitive browser tasks and workflows
- Scraping data from websites programmatically
- Testing across Chrome, Firefox, Safari, and Edge browsers
- Taking screenshots for visual regression testing
- Testing authentication and login flows
- Filling and submitting forms automatically
- Testing responsive designs across different viewports
- Simulating user interactions (clicks, typing, navigation)
- Testing file uploads and downloads
- Capturing network requests and responses
- Testing Single Page Applications (SPAs)
- Any browser automation requiring reliable cross-browser support

📊 All 50 Skills Optimized

Core Development (20 skills)

✓ test-driven-development · code-review · debugging-systematic
✓ git-workflow · refactoring · performance-optimization
✓ security-patterns · api-design · database-design
✓ frontend-component-patterns · backend-service-patterns · e2e-testing
✓ ci-cd-pipeline · documentation-generation · error-handling-patterns
✓ accessibility-wcag · standards-enforcement · data-migration
✓ monitoring-observability · incident-response

Framework-Specific (10 skills)

✓ nextjs-app-router · nextjs-server-components · neondb-serverless
✓ convex-backend · convex-realtime · shadcn-ui-components
✓ tailwind-design-system · react-server-actions · typescript-strict
✓ vercel-deployment

Developer Tools (10 skills)

✓ mcp-builder · artifacts-builder · playwright-automation
✓ root-cause-tracing · brainstorming · changelog-generator
✓ content-research-writer · document-processing-docx
✓ document-processing-pdf · document-processing-xlsx

Productivity (10 skills)

✓ file-organizer · invoice-organizer · domain-name-brainstormer
✓ competitive-research · brand-guidelines · canvas-design
✓ meeting-insights-analyzer · kaizen-continuous-improvement
✓ threat-hunting · skill-creator

🎯 Benefits

For Users

✅ Dramatically improved skill discoverability
✅ Crystal-clear activation criteria with 12+ examples each
✅ Professional, consistent quality across entire library
✅ Comprehensive usage examples for every scenario
✅ Better AI skill activation based on detailed descriptions

For Skills Library

✅ All 50 skills optimized to maximum quality
✅ 100 SKILL.md files updated (50 Factory + 50 Claude Code)
✅ Consistent structure using skill-creator patterns
✅ Production-ready, professional quality throughout
✅ Ready for enterprise and professional use

📦 Installation

Fresh Install

curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh | bash

Upgrade Existing Installation

cd your-project
curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh | bash

📝 Files Changed

• 100 SKILL.md files updated (50 per platform)
• All skills synced between Factory Droid CLI and Claude Code
• CHANGELOG.md updated with comprehensive examples

---

Full Changelog: https://github.com/korallis/Droidz/blob/main/CHANGELOG.md

This release represents the culmination of our skills optimization work - every skill in Droidz is now professionally documented with comprehensive descriptions and extensive usage examples! 🎉

v4.7.1 - improve-skills Uses skill-creator

🎯 Better Skill Improvements

/improve-skills now leverages the skill-creator skill for higher quality improvements.

What Changed

improve-skills now uses skill-creator throughout:

• ✅ skill-creator templates for structure
• ✅ skill-creator patterns for descriptions
• ✅ skill-creator best practices applied
• ✅ Consistent quality improvements

Benefits

• Higher quality skill improvements
• Consistent structure across skills
• Leverages existing expertise
• Better activation criteria

Usage

> /improve-skills
AI: Using skill-creator to improve skills...
✅ All skills improved with skill-creator patterns

No user action required - skill-creator automatically used!

---

Changelog: v4.7.0...v4.7.1

v4.7.0 - Standards-Shaper Updates Existing Standards

🔄 Intelligent Standards Updates

standards-shaper now intelligently updates existing standards to match your project's actual tech stack.

The Problem

Generic standards didn't match projects:

• ❌ Examples for wrong frameworks
• ❌ Manual updates required
• ❌ Standards became stale
• ❌ Not project-specific

The Solution

Two Intelligent Modes:

1. Create Mode - New standards with project-specific examples
2. Update Mode ⭐ NEW - Updates existing standards to match current stack

How It Works

1. Detects existing standards
2. Analyzes project tech stack
3. Updates examples to match
4. Preserves customizations

What Gets Updated

• error-handling, testing, components
• styling, state-management, api-design
• database, authentication, and more

All with project-specific patterns!

Usage

# After framework upgrade
> /standards-shaper
AI: Updating standards...
AI: Detected React 19, Fastify, Prisma
✅ Updated standards to match stack

Benefits

• ✅ Standards match your frameworks
• ✅ Re-run after upgrades
• ✅ Project-specific examples
• ✅ No breaking changes

Installation

curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh | bash

---

Full Changelog: v4.6.0...v4.7.0

v4.6.0 - TodoWrite Progress Tracking for Factory Droids

👁️ Subagent Visibility Enhancement

Factory.ai subagents now show real-time progress using TodoWrite, solving the "check on subagent" problem.

The Problem

Before v4.6.0, subagents in Factory.ai ran invisibly:

• ❌ No feedback during long operations
• ❌ Couldn't see what subagents were doing
• ❌ Had to manually "check on subagent"
• ❌ Poor UX for multi-step workflows

The Solution

Added TodoWrite progress tracking to 5 core Factory droids.

What's New

5 Droids with Real-Time Progress:

1. product-planner - Tracks: requirements, mission, roadmap, tech stack
2. spec-shaper - Tracks: analysis, questions, responses (X/Y answered)
3. spec-writer - Tracks: reading, analyzing, writing (section X/Y)
4. tasks-list-creator - Tracks: analyzing, breaking into groups (X created)
5. implementer - Tracks: planning, implementing (X/Y complete), testing

User Experience

Before:

> /write-spec
AI: Creating specification...
[Long silence with no updates]

After:

> /write-spec
AI: Creating specification...

Live Progress:
✅ Reading requirements (92 Q&A)
✅ Analyzing and structuring (12 sections)
🔄 Writing specification (section 7/12)
⏸️ Validating completeness

How It Works

Subagents use TodoWrite to update progress:

• Create todos at start
• Update as work progresses
• Mark complete when done

Updates appear in main session automatically!

Benefits

• ✅ Real-time subagent visibility
• ✅ No more "check on subagent" needed
• ✅ Better workflow understanding
• ✅ Especially helpful for long operations
• ✅ Zero breaking changes

Platform Strategy

Factory Droid CLI: Progress tracking added
Claude Code: Unchanged (avoids conflicts)

Installation

curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh | bash

Progress tracking works automatically!

---

Full Changelog: v4.5.0...v4.6.0

v4.5.0 - Optional MCP Research Integration

🔬 Intelligent Research Enhancement

All agents/droids now automatically leverage Exa and Ref MCPs when available, with graceful degradation when not installed.

What's New

5 Core Agents Enhanced with Research Capabilities:

1. product-planner

   • Similar product architectures and patterns
   • Technology stack best practices
   • Market trends and competitor analysis

2. spec-shaper

   • Technical architecture patterns
   • Similar feature implementations
   • Framework-specific best practices

3. spec-writer

   • API design patterns and conventions
   • Database schema best practices
   • Security implementation patterns

4. tasks-list-creator

   • Implementation patterns for similar features
   • Task breakdown strategies
   • Common pitfalls and dependencies

5. implementer

   • Code examples and patterns
   • Framework-specific best practices
   • Common solutions to challenges

How It Works

Graceful Integration:

• Instructions only (no hard dependencies)
• Automatic usage when MCPs available
• Graceful continuation without MCPs
• No errors or failures
• Enhanced quality when available

Benefits

With Exa/Ref MCPs:

• 📊 Market research for product planning
• 🏗️ Current best practices for specs
• 💻 Proven code examples for implementation

Without MCPs:

• ✅ All agents work normally
• ✅ General knowledge and best practices
• ✅ No setup required

Usage

# Install/upgrade Droidz
curl -fsSL https://raw.githubusercontent.com/korallis/Droidz/main/install.sh | bash

# Agents automatically use MCPs if installed
> /plan-product
> /shape-spec
> /write-spec
> /create-tasks

Files Modified

10 files updated (5 agents × 2 platforms):

• product-planner, spec-shaper, spec-writer
• tasks-list-creator, implementer
• Both Factory Droid CLI and Claude Code

---

Full Changelog: v4.4.1...v4.5.0