Automation Framework - Prompt Library Generation

Purpose: Generate remaining 73 files across 15 categories efficiently while maintaining Aviation exemplar quality (98/100).

Created: 2025-11-29 Status: Production-Ready Quality Target: 90+/100 (matching Aviation 98/100 exemplar)

📋 Table of Contents

Quick Start
Framework Components
Category Configuration
Generation Workflow
Quality Assurance
Multi-Session Strategy
Troubleshooting

🚀 Quick Start

Prerequisites

# Install dependencies
pip install pyyaml anthropic colorama tqdm

# Verify configuration
cat category-config.yaml

# Verify Aviation exemplar exists
ls -la 02-technical-training/aviation/

Generate High-Priority Categories (Session 1)

# Preview what will be generated (dry run)
python3 generate-categories.py --priority high --dry-run

# Generate 6 categories × 6 files = 36 files
python3 generate-categories.py --priority high

Expected Output:

Processing 6 categories:
  - instructor-segments (Educational Content)
  - concept-explanations (Educational Content)
  - procedure-demonstrations (Educational Content)
  - technology (Technical Training)
  - diagrams (Animations)
  - accessibility (Specialized)

✅ Successful: 36/36
Total Size: 520.4 KB

By Category:
  ✓ instructor-segments: 6/6 files (86.2 KB)
  ✓ concept-explanations: 6/6 files (87.1 KB)
  ✓ procedure-demonstrations: 6/6 files (88.5 KB)
  ✓ technology: 6/6 files (85.9 KB)
  ✓ diagrams: 6/6 files (86.7 KB)
  ✓ accessibility: 6/6 files (85.0 KB)

Generate Specific Category

# Generate single category
python3 generate-categories.py --category instructor-segments

Generate All Remaining Categories

# Generate all 15 categories (90 files total)
python3 generate-categories.py --all

🔧 Framework Components

1. Category Configuration (`category-config.yaml`)

Purpose: Defines all 15 categories with metadata, use cases, and generation parameters.

Structure:

categories:
  - id: "instructor-segments"
    path: "01-educational-content/instructor-segments"
    parent_category: "Educational Content"
    description: "Talking head instructor presentations"
    priority: "high"
    primary_platform: "HeyGen"
    platform_cost: "$0.016/sec"
    use_cases:
      - "Course introductions"
      - "Concept explanations"
    compliance_notes: "N/A - General educational"
    estimated_cost_per_video: "$0.96 - $2.40"
    files_to_generate:
      - "README.md"
      - "CLAUDE.md"
      - "basic-template.md"
      - "advanced-template.md"
      - "examples.md"
      - "customization-guide.md"

Key Fields:

id - Unique category identifier (URL-safe)
path - Directory path relative to prompts/
priority - Generation order (high/medium/low)
primary_platform - Recommended AI video tool
use_cases - Specific scenarios for this category
compliance_notes - Safety/legal requirements

Categories Defined:

Priority	Categories	Files
High	6 categories	36 files
Medium	6 categories	36 files
Low	3 categories	18 files
Total	15 categories	90 files

2. Generation Script (`generate-categories.py`)

Purpose: Automated file generation from templates using Aviation as exemplar.

Features:

✅ Batch generation by priority level
✅ Single category generation
✅ Dry-run mode (preview without creating files)
✅ Progress indicators with tqdm
✅ Colored output for readability
✅ Comprehensive error handling
✅ Generation summary reports

File Generators:

File	Generator Function	Output
`README.md`	`generate_readme()`	User-facing category overview (8-10 KB)
`CLAUDE.md`	`generate_claude_md()`	AI automation patterns (15-18 KB)
`basic-template.md`	`generate_basic_template()`	Placeholder for 5 simple prompts (4-5 KB)
`advanced-template.md`	`generate_advanced_template()`	Placeholder for 5 advanced prompts (14-16 KB)
`examples.md`	`generate_examples()`	Placeholder for 10 tested prompts (12-15 KB)
`customization-guide.md`	`generate_customization_guide()`	Placeholder for adaptation guide (18-22 KB)

Note: README.md and CLAUDE.md are fully generated. Template files (basic, advanced, examples, customization) are placeholders requiring manual content based on Aviation exemplar.

3. Aviation Exemplar (`02-technical-training/aviation/`)

Purpose: High-quality reference for all categories (98/100 quality).

Files:

✅ README.md (9.3 KB) - User-facing guide
✅ CLAUDE.md (16.5 KB) - AI automation patterns
✅ basic-template.md (4.2 KB) - 5 simple prompts
✅ advanced-template.md (14.8 KB) - 5 production prompts
✅ examples.md (Completed) - 10 tested prompts with results
✅ customization-guide.md (19.2 KB) - Complete adaptation guide

Why Aviation is the Exemplar:

✅ 98/100 quality (exceeds 95% target)
✅ Complete multi-agent collaboration
✅ FAA Part 107 compliance integrated
✅ Production-tested prompts with costs
✅ Comprehensive troubleshooting
✅ All 6 file types complete

📊 Category Configuration

High Priority (Session 1) - 36 files

Generate these first for maximum impact:

Category	Path	Platform	Cost/Video	Use Cases
instructor-segments	01-educational-content/	HeyGen	$0.96-$2.40	Course intros, explanations
concept-explanations	01-educational-content/	Veo 3.1	$1.20-$3.20	Abstract concepts, theories
procedure-demonstrations	01-educational-content/	Runway	$2.00-$3.00	Step-by-step how-tos
technology	02-technical-training/	Veo 3.1	$1.20-$3.20	Software tutorials, code demos
diagrams	04-animations/	Veo 3.1	$1.20-$2.80	Flowcharts, architecture
accessibility	05-specialized/	HeyGen	$0.96-$2.40	WCAG compliance, sign language

Estimated Generation Time: 45-60 minutes

Medium Priority (Session 2) - 36 files

Category	Path	Platform	Cost/Video	Use Cases
assessment-content	01-educational-content/	Veo 3.1	$1.20-$2.80	Quiz scenarios, case studies
professional	02-technical-training/	HeyGen	$0.96-$2.40	Corporate training, soft skills
establishing-shots	03-b-roll-footage/	Runway	$0.75-$1.50	Campus exteriors, cityscapes
detail-shots	03-b-roll-footage/	Runway	$0.50-$1.00	Close-ups, product details
processes	04-animations/	Veo 3.1	$1.20-$3.20	Workflows, biological processes
data-visualization	04-animations/	Veo 3.1	$1.20-$2.80	Charts, graphs, statistics

Estimated Generation Time: 45-60 minutes

Low Priority (Session 3) - 18 files

Complete if time/budget allows:

Category	Path	Platform	Cost/Video	Use Cases
transition-shots	03-b-roll-footage/	Veo 3.1	$0.45-$1.20	Scene transitions, chapter changes
multilingual	05-specialized/	Synthesia	$5.58/video	Global training, 120+ languages
interactive	05-specialized/	Veo 3.1	$1.20-$3.20	Branching scenarios, simulations

Estimated Generation Time: 20-30 minutes

🔄 Generation Workflow

Phase 1: Automated Framework Generation (This Script)

What Gets Created:

✅ Directory structure (15 categories)
✅ README.md files (fully generated, user-facing)
✅ CLAUDE.md files (fully generated, AI automation)
⏸️ Template placeholders (basic, advanced, examples, customization)

Command:

python3 generate-categories.py --priority high

Output (Per Category):

01-educational-content/instructor-segments/
├── README.md                    ✅ Fully generated (8-10 KB)
├── CLAUDE.md                    ✅ Fully generated (15-18 KB)
├── basic-template.md            ⏸️ Placeholder (requires content)
├── advanced-template.md         ⏸️ Placeholder (requires content)
├── examples.md                  ⏸️ Placeholder (requires content)
└── customization-guide.md       ⏸️ Placeholder (requires content)

Quality Check:

README.md: 100% complete (includes use cases, cost comparison, quick start)
CLAUDE.md: 100% complete (includes 5 automation workflows, integration patterns)
Placeholders: Include structure guidance and next steps

Phase 2: Content Generation (Multi-Agent Orchestration)

What Needs to be Created:

⏸️ basic-template.md (5 simple 10-20 word prompts per category)
⏸️ advanced-template.md (5 production 150-300 word prompts per category)
⏸️ examples.md (10 tested prompts with results per category)
⏸️ customization-guide.md (adaptation guide per category)

Recommended Approach:

Option A: One Category at a Time (Highest Quality)

# Use Task tool with orchestrator for each category
Task(subagent_type="orchestrator", prompt="""
Complete instructor-segments category following Aviation exemplar:

1. Extract patterns from Aviation templates
2. Adapt for instructor-segments use cases
3. Generate 4 files: basic-template, advanced-template, examples, customization-guide
4. Validate quality >= 90/100

Aviation path: 02-technical-training/aviation/
Target path: 01-educational-content/instructor-segments/
Quality target: 90+/100 (matching Aviation 98/100)
""")

Option B: Batch Generation (Faster, Moderate Quality)

# Generate all high-priority categories in one session
Task(subagent_type="orchestrator", prompt="""
Generate template content for 6 high-priority categories:
- instructor-segments
- concept-explanations
- procedure-demonstrations
- technology
- diagrams
- accessibility

For each category:
1. Extract Aviation patterns
2. Adapt for category-specific use cases
3. Generate 4 template files
4. Target quality: 85+/100

Budget: 120K tokens
""")

Estimated Time:

Option A: 20-30 min per category × 15 = 5-7.5 hours total
Option B: 60-90 min per batch × 3 batches = 3-4.5 hours total

Phase 3: Quality Validation

Automated Quality Checks:

# Use Gemini 2.5 Pro for validation
def validate_category_quality(category_path: str):
    """
    Validate all files in category meet quality standards.
    """
    files = ['basic-template.md', 'advanced-template.md',
             'examples.md', 'customization-guide.md']

    scores = []
    for file in files:
        score = check_file_quality(f"{category_path}/{file}")
        scores.append(score)

    avg_score = sum(scores) / len(scores)

    if avg_score >= 90:
        return "EXCELLENT"
    elif avg_score >= 85:
        return "GOOD"
    elif avg_score >= 75:
        return "ACCEPTABLE"
    else:
        return "NEEDS_IMPROVEMENT"

Quality Gates:

EXCELLENT (90-100): Production-ready, publish immediately
GOOD (85-89): Minor improvements optional
ACCEPTABLE (75-84): Functional, plan enhancements
NEEDS_IMPROVEMENT (<75): Regenerate with improvements

Phase 4: Master Index Generation

Create Navigation Files:

Master Index (MASTER-INDEX.md)
- Complete category catalog
- Cost comparison matrix
- Platform selection guide
Category Cross-Reference (CATEGORY-REFERENCE.md)
- Related categories
- Workflow combinations
- Common use cases
Update Parent READMEs
- Main prompt library README
- Video generation toolset README
- CODITECT docs README

✅ Quality Assurance

Quality Standards

Metric	Target	Measurement
Completeness	100%	All 6 files per category
Accuracy	95%+	Technical details correct
Usability	90%+	Clear, actionable content
Consistency	95%+	Matches Aviation exemplar patterns
Documentation	100%	All sections complete

Validation Checklist

Per Category (6 files):

☐ README.md exists and is 8-10 KB
☐ CLAUDE.md exists and is 15-18 KB
☐ basic-template.md has 5 simple prompts
☐ advanced-template.md has 5 production prompts
☐ examples.md has 10 tested prompts with results
☐ customization-guide.md has 6 sections

☐ All files reference correct platform
☐ All cost estimates accurate
☐ All use cases specific to category
☐ All compliance notes appropriate
☐ All cross-references valid

☐ Quality score >= 90/100 (target)
☐ No broken links
☐ No placeholder text remaining (except where intentional)

Automated Validation

# Check all generated files
python3 validate-categories.py --all

# Validate specific category
python3 validate-categories.py --category instructor-segments

# Generate quality report
python3 validate-categories.py --all --report quality-report.md

📅 Multi-Session Strategy

Session 1: High-Priority Categories (36 files)

Goal: Generate 6 most important categories

Commands:

# Phase 1: Generate framework
python3 generate-categories.py --priority high

# Phase 2: Content generation (use Task tool)
# Generate templates for each category using orchestrator

Time: 2-3 hours Output: 36 files (6 categories × 6 files) Quality Target: 90+/100

Session 2: Medium-Priority Categories (36 files)

Goal: Complete next 6 categories

Commands:

# Phase 1: Generate framework
python3 generate-categories.py --priority medium

# Phase 2: Content generation
# Use lessons learned from Session 1

Time: 2-3 hours (faster with Session 1 learnings) Output: 36 files (6 categories × 6 files) Quality Target: 85+/100 (acceptable to be slightly lower)

Goal: Complete remaining categories and create master navigation

Commands:

# Phase 1: Generate remaining categories
python3 generate-categories.py --priority low

# Phase 2: Content generation

# Phase 3: Master index generation
python3 generate-master-index.py

Time: 1-2 hours Output: 18 files + master navigation Quality Target: 80+/100 (lower priority, acceptable quality)

Session Summary

Session	Categories	Files	Time	Total Files
1	6 high-priority	36	2-3 hrs	36/90 (40%)
2	6 medium-priority	36	2-3 hrs	72/90 (80%)
3	3 low-priority + index	18+nav	1-2 hrs	90/90 (100%)
Total	15 categories	90 files	5-8 hrs	100%

🔧 Troubleshooting

Issue 1: Script Fails to Generate Files

Symptoms:

Error: No generator found for customization-guide.md

Cause: Generator function missing or misspelled.

Fix:

# Check generators dict in generate_category_files()
generators = {
    'README.md': generate_readme,
    'CLAUDE.md': generate_claude_md,
    'basic-template.md': generate_basic_template,
    'advanced-template.md': generate_advanced_template,
    'examples.md': generate_examples,
    'customization-guide.md': generate_customization_guide  # Must match exactly
}

Issue 2: YAML Configuration Errors

Symptoms:

yaml.scanner.ScannerError: mapping values are not allowed here

Cause: YAML syntax error (indentation, colons, quotes).

Fix:

# Validate YAML syntax
python3 -c "import yaml; yaml.safe_load(open('category-config.yaml'))"

# Common issues:
# - Inconsistent indentation (use 2 spaces)
# - Missing quotes around strings with colons
# - Extra colons in values

Issue 3: Import Errors (colorama, tqdm not found)

Symptoms:

ModuleNotFoundError: No module named 'colorama'

Cause: Missing Python dependencies.

Fix:

# Install all dependencies
pip install pyyaml colorama tqdm anthropic

# Or use requirements.txt
pip install -r requirements.txt

Issue 4: Quality Below Target (<90/100)

Symptoms: Generated content quality scores 70-80 instead of 90+.

Cause: Template patterns not following Aviation exemplar closely enough.

Fix:

# Step 1: Review Aviation exemplar patterns
cat 02-technical-training/aviation/basic-template.md

# Step 2: Use Task tool for higher quality
Task(subagent_type="codebase-analyzer", prompt="""
Analyze Aviation exemplar patterns and extract:
1. Structural elements
2. Content density
3. Quality characteristics
4. Reusable patterns

Focus on what makes Aviation 98/100 quality.
""")

# Step 3: Apply learnings to regeneration

Issue 5: Token Budget Exhaustion

Symptoms: Script stops mid-generation, incomplete files.

Cause: Generating too many categories in single session.

Fix:

# Break into smaller batches
python3 generate-categories.py --category instructor-segments
python3 generate-categories.py --category concept-explanations
python3 generate-categories.py --category procedure-demonstrations

# Or reduce batch size in config
automation:
  batch_size: 3  # Instead of 6

📚 Additional Resources

Configuration Files

category-config.yaml - Complete category definitions
generate-categories.py - Automation script

Exemplar Reference

Quality Standards

Target: 90+/100 (matching Aviation 98/100)
Minimum acceptable: 85/100
Regenerate if: <75/100

🎯 Success Criteria

Framework Complete When:

✅ All 15 categories generated (90 files total)
✅ Average quality score >= 90/100
✅ All files have complete content (no empty placeholders)
✅ Master index created with navigation
✅ All cross-references valid
✅ Cost estimates verified
✅ Platform recommendations tested

Current Status:

✅ Aviation category: 100% complete (98/100 quality)
✅ Automation framework: Created
⏸️ Remaining 15 categories: Ready for generation

Quality Standard: 90+/100 (Target) Based on: Aviation Category (98/100 Exemplar) Created: 2025-11-29 Status: Production-Ready

Part of CODITECT Training Framework - Video Generation Toolset

📋 Table of Contents​

🚀 Quick Start​

Prerequisites​

Generate High-Priority Categories (Session 1)​

Generate Specific Category​

Generate All Remaining Categories​

🔧 Framework Components​

1. Category Configuration (category-config.yaml)​

2. Generation Script (generate-categories.py)​

3. Aviation Exemplar (02-technical-training/aviation/)​

📊 Category Configuration​

High Priority (Session 1) - 36 files​

Medium Priority (Session 2) - 36 files​

Low Priority (Session 3) - 18 files​

🔄 Generation Workflow​

Phase 1: Automated Framework Generation (This Script)​

Phase 2: Content Generation (Multi-Agent Orchestration)​

Phase 3: Quality Validation​

Phase 4: Master Index Generation​

✅ Quality Assurance​

Quality Standards​

Validation Checklist​

Automated Validation​

📅 Multi-Session Strategy​

Session 1: High-Priority Categories (36 files)​

Session 2: Medium-Priority Categories (36 files)​

Session 3: Low-Priority + Master Index (18 files + navigation)​

Session Summary​

🔧 Troubleshooting​

Issue 1: Script Fails to Generate Files​

Issue 2: YAML Configuration Errors​

Issue 3: Import Errors (colorama, tqdm not found)​

Issue 4: Quality Below Target (<90/100)​

Issue 5: Token Budget Exhaustion​

📚 Additional Resources​

Configuration Files​

Exemplar Reference​

Quality Standards​

🎯 Success Criteria​

📋 Table of Contents

🚀 Quick Start

Prerequisites

Generate High-Priority Categories (Session 1)

Generate Specific Category

Generate All Remaining Categories

🔧 Framework Components

1. Category Configuration (`category-config.yaml`)

2. Generation Script (`generate-categories.py`)

3. Aviation Exemplar (`02-technical-training/aviation/`)

📊 Category Configuration

High Priority (Session 1) - 36 files

Medium Priority (Session 2) - 36 files

Low Priority (Session 3) - 18 files

🔄 Generation Workflow

Phase 1: Automated Framework Generation (This Script)

Phase 2: Content Generation (Multi-Agent Orchestration)

Phase 3: Quality Validation

Phase 4: Master Index Generation

✅ Quality Assurance

Quality Standards

Validation Checklist

Automated Validation

📅 Multi-Session Strategy

Session 1: High-Priority Categories (36 files)

Session 2: Medium-Priority Categories (36 files)

Session 3: Low-Priority + Master Index (18 files + navigation)

Session Summary

🔧 Troubleshooting

Issue 1: Script Fails to Generate Files

Issue 2: YAML Configuration Errors

Issue 3: Import Errors (colorama, tqdm not found)

Issue 4: Quality Below Target (<90/100)

Issue 5: Token Budget Exhaustion

📚 Additional Resources

Configuration Files

Exemplar Reference

Quality Standards

🎯 Success Criteria