# Enhancer Promptforge Advanced Innovation

> Use when developing advanced prompt techniques, pushing the boundaries of prompt effectiveness, and creating innovative approaches beyond established best practices. This skill transforms proven techniques into advanced, optimized methodologies.

- Skill: `construct-ai-primary/enhancer-promptforge-advanced-innovation` (Agent Skill)
- Install (CLI): `npx skillmds@latest add construct-ai-primary/enhancer-promptforge-advanced-innovation`
- Raw SKILL.md: https://api.skillmd.com/api/skills/construct-ai-primary/enhancer-promptforge-advanced-innovation/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: Construct-AI-primary (https://skillmd.com/u/construct-ai-primary)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/construct-ai-primary/enhancer-promptforge-advanced-innovation

---


# Enhancer - Advanced Techniques & Innovation

## Overview
Enhancer develops advanced prompt techniques for PromptForge AI. Once Explorer identifies and Experimenter validates a technique, Enhancer takes it further—optimizing, combining, and innovating to create techniques that are better than anything published.

## When to Use
- When taking a proven technique and pushing it beyond published limits
- When combining multiple techniques into a novel hybrid approach
- When developing proprietary prompt engineering methods
- When optimizing techniques for specific models, domains, or use cases
- When creating next-generation approaches that don't exist in literature yet
- **Don't use when:** Researching new techniques (use explorer) or validating claims (use experimenter)

## Core Procedures

### Step 1: Identify Enhancement Opportunities
Start with proven techniques that have room for improvement:
- What are the limitations of the current best approach?
- Where do published techniques fall short in practice?
- What combinations of techniques haven't been tried?
- What model-specific optimizations are possible?

### Step 2: Design Advanced Technique
Apply enhancement patterns:
- **Composition:** Combine 2+ techniques (e.g., Chain-of-Thought + Self-Consistency)
- **Iteration:** Add refinement loops to single-pass techniques
- **Adaptation:** Make static techniques dynamic based on input characteristics
- **Meta-Learning:** Create techniques that learn from their own outputs
- **Constraint Engineering:** Add specific constraints to guide better outputs

### Step 3: Prototype and Iterate
1. Build minimum viable version of the advanced technique
2. Test against baseline (current best practice)
3. Measure improvement across multiple dimensions (quality, cost, speed, reliability)
4. Identify failure modes and edge cases
5. Refine based on results
6. Repeat until improvement is significant and consistent

### Step 4: Validate Rigorously
Before declaring a technique "advanced":
- [ ] Tested on 50+ diverse inputs
- [ ] Tested across 2+ model families
- [ ] Tested in real-world conditions (not just controlled experiments)
- [ ] Improvement is statistically significant (p < 0.05)
- [ ] Improvement is practically significant (effect size > 0.5)
- [ ] No regression on any important dimension

### Step 5: Document and Transfer
```
ADVANCED TECHNIQUE DOCUMENT
===========================
Technique Name: [descriptive name]
Base Techniques: [what this builds on]
Innovation: [what's new and why it's better]
When to Use: [specific conditions where this excels]
When Not to Use: [conditions where this fails or is overkill]
Implementation: [step-by-step instructions]
Performance: [measured improvement over baseline]
Model Compatibility: [which models support this]
Transfer To: [experimenter for formal testing, archivist for knowledge base]
```

## Success Metrics
- Advanced techniques developed: >=2 per quarter
- Improvement over baseline: >=15% quality improvement on average
- Adoption rate: >=70% of advanced techniques adopted into standard practice
- Innovation pipeline: >=5 techniques in development at any time

## Error Handling
- **Error:** Advanced technique is too complex for practical use
  **Response:** Simplify while preserving core innovation, document complexity trade-off
- **Error:** Technique works on one model but fails on others
  **Response:** Document model-specific limitations, investigate why it fails, adapt if possible
- **Error:** Improvement is marginal (<5%)
  **Response:** Not worth the added complexity—abandon or radically redesign

## Cross-Team Integration
**Related Skills:** explorer-promptforge-new-techniques, experimenter-promptforge-hypothesis-testing, specialist-promptforge-domain-adaptation, innovation-evaluation
**Used By:** Explorer, Experimenter, Specialist, Sage, Tuner