This repository contains a collection of my most advanced prompts and meta-systems designed for prompt engineering, debugging, evaluation, and specialized AI orchestration tasks.

New to prompt engineering? Start with the Dual Path Primer to understand your requirements, then use the Ultimate Prompt Evaluator to refine your prompts.

Building complex prompts from scratch in CLI/terminal environments? Use the Prompt Creation Trilogy for systematic analysis, generation, and improvement. Designed specifically for Claude Code, Codex, and similar terminal-based AI development workflows.

Debugging AI systems or code? Try the Adaptive Debug Protocol for systematic problem-solving.

Working with multi-agent AI systems? Use the Multi-Agent Orchestration Evaluator for comprehensive system analysis.

Need to add features to existing prompts with strict character limits? Use PROMPTGRAFT for surgical precision editing with budget enforcement and quality gates.

Need structured development with TDD and quality gates? Use Journeyman for 5-phase disciplined implementation from Blueprint to Verification.

Location: prompt_primer_and_evaluator/UPE/

A comprehensive meta-prompt that provides 34-point evaluation of any prompt you submit. Features instant analysis, iterative refinement, and the ability to transform concepts into complete prompts.

Key Features:

• 34-criteria evaluation framework
• Iterative "re-evaluate" command for continuous improvement
• Concept-to-prompt transformation
• Platform-aware optimizations (Claude, ChatGPT, Gemini)

Best For: Evaluating and refining existing prompts, quality assurance, prompt optimization

Location: prompt_primer_and_evaluator/dual_path_primer/

An AI meta-prompt orchestrator that refuses to proceed until it achieves 100% understanding of your request. Creates either perfectly contextualized conversations or highly optimized prompts.

Key Features:

• Active understanding pursuit with gap detection
• Structured "Readiness Report Table"
• Internal self-verification systems
• Dual output options (conversation or prompt)

Best For: Complex requirements gathering, creating context-rich prompts, setting up expert AI conversations

Perfect Workflow: Dual Path Primer → UPE → Deployment

Location: prompt-creation-trilogy/

A three-stage systematic approach: Analyze → Generate → Improve

Deep problem analysis using multiple frameworks (Ishikawa, Five Whys, OODA loops) before writing any prompts.

Autonomous prompt architect that builds prompts iteratively, scoring its own work through 6 mandatory loops.

Takes any existing prompt and systematically enhances it to achieve 90%+ quality scores using domain-aware improvements.

Best For: Complex problem-solving, systematic prompt development, professional prompt engineering workflows

Location: prompt-adaptive-debug-protocol/

Military-grade debugging system using OODA loops (Observe, Orient, Decide, Act) with escalating cognitive intensity for systematic problem-solving.

Key Features:

• Automatic thinking level escalation (think → megathink → ultrathink)
• Structured 4-loop maximum with forced breakthrough thinking
• Comprehensive debug logging (debug_loop.md files)
• Built-in problem classification and framework selection

Best For: Complex debugging challenges, systematic troubleshooting, breaking out of failed solution loops

Location: prompt-multi-agent-orchestration.evaluator/

Comprehensive evaluator for multi-agent AI systems with 40+ quality criteria, modern framework support, and cost optimization analysis.

Key Features:

• 40-point evaluation framework covering architecture, performance, security
• Framework-aware analysis (AutoGen, LangGraph, CrewAI, etc.)
• Cost-benefit analysis with ROI projections
• Implementation roadmaps and compliance assessment

Best For: Enterprise AI systems, multi-agent architectures, production system optimization

Location: promptgraft/

A pure LLM pipeline for surgical prompt editing. Six AI specialists coordinate to add features to prompts with strict character budget enforcement and quality gates. 95% success rate vs 40% manual editing.

Key Features:

• 6-specialist pipeline (Architect → Surgeon → Auditor → Executor → Inspector → Chronicler)
• Strict character budget enforcement (never exceeded)
• Three quality gates (pre-execution, post-execution, budget verification)
• Complete traceability and version history
• run-all command for automatic hands-free execution

Best For: Production prompts with character limits, evolving prompt systems (v1 → v20), team collaboration, API prompts with token budgets

Location: journeyman/

A structured 5-phase development methodology that ensures every task is completed with discipline, quality, and traceability. TDD mandatory, quality gates enforced, full documentation.

Key Features:

• 5 phases: Blueprint → Foundation → Assembly → Finishing → Verification
• TDD mandatory (tests before implementation, always)
• Quality gates (can't advance until criteria met)
• JOURNEY.md tracking (single source of truth for decisions)
• Simple Path (6 steps) for tasks < 2 days
• Full Protocol (5 phases) for complex 2+ day projects
• Framework library with implementation playbooks

Best For: Features that must work the first time, refactoring with zero regressions, team collaboration, complex systems (5+ files), learning TDD discipline

1. Dual Path Primer - Understand requirements deeply
2. Ultimate Prompt Evaluator - Evaluate and refine the result
3. Deploy - Use the optimized prompt

1. Prompt Creation Trilogy (Analyzer) - Understand the problem
2. Prompt Creation Trilogy (Generator) - Build targeted prompt
3. Prompt Creation Trilogy (Improver) - Polish to perfection

1. Adaptive Debug Protocol - Systematic problem analysis
2. Ultimate Prompt Evaluator - Refine any prompts used in solutions

1. Multi-Agent Orchestration Evaluator - Assess current architecture
2. Dual Path Primer - Plan improvements
3. Ultimate Prompt Evaluator - Optimize individual agent prompts

1. PROMPTGRAFT - Add features with character budget control
2. Ultimate Prompt Evaluator - Validate the enhanced prompt
3. Deploy - Use the surgically modified prompt

1. Journeyman - Follow 5-phase protocol (Blueprint → Verification)
2. Ultimate Prompt Evaluator - Validate any prompts created during development
3. Deploy - Ship with confidence knowing all gates passed

1. Choose your system based on your needs (see "When to Use" table above)
2. Navigate to the appropriate directory
3. Read the system-specific README for detailed instructions
4. Copy the prompt into your preferred AI platform
5. Follow the system's workflow for optimal results

I work relentlessly on prompt engineering and am always focused on building many prompts while staying constantly involved in this field. This repository will continue to serve as a place where I share refined prompts and new additions as they develop.

Found these prompts valuable? Please consider giving this repository a ⭐ star to help others discover these professional-grade prompt engineering tools!