name: fabric context: fork model: sonnet description: Intelligent pattern selection for Fabric CLI. Automatically selects the right pattern from 242+ specialized prompts based on your intent - threat modeling, analysis, summarization, content creation, extraction, and more. USE WHEN processing content, analyzing data, creating summaries, threat modeling, or transforming text.

Fabric Skill

Setup Check - Fabric Repository

IMPORTANT: Before using this skill, verify the Fabric repository is available:

# Check if Fabric exists
fabric --version && \
test -f ~/.config/fabric/patterns/pattern_explanations.md && echo 'Fabric is properly installed and configured'

If Fabric is not installed, alert the user, make sure it is installed.

When to Activate This Skill

Primary Use Cases:

"Create a threat model for..."
"Summarize this article/video/paper..."
"Extract wisdom/insights from..."
"Analyze this [code/malware/claims/debate]..."
"Improve my writing/code/prompt..."
"Create a [visualization/summary/report]..."
"Rate/review/judge this content..."

The Goal: Select the RIGHT pattern from 242+ available patterns based on what you're trying to accomplish.

🎯 Pattern Selection Strategy

When a user requests Fabric processing, follow this decision tree:

1. Identify Intent Category

Threat Modeling & Security:

Threat model → create_threat_model or create_stride_threat_model
Threat scenarios → create_threat_scenarios
Security update → create_security_update
Security rules → create_sigma_rules, write_nuclei_template_rule, write_semgrep_rule
Threat analysis → analyze_threat_report, analyze_threat_report_trends

Summarization:

General summary → summarize
5-sentence summary → create_5_sentence_summary
Micro summary → create_micro_summary or summarize_micro
Meeting → summarize_meeting
Paper/research → summarize_paper
Video/YouTube → youtube_summary
Newsletter → summarize_newsletter
Code changes → summarize_git_changes or summarize_git_diff

Wisdom Extraction:

General wisdom → extract_wisdom
Article wisdom → extract_article_wisdom
Book ideas → extract_book_ideas
Insights → extract_insights or extract_insights_dm
Main idea → extract_main_idea
Recommendations → extract_recommendations
Controversial ideas → extract_controversial_ideas

Analysis:

Malware → analyze_malware
Code → analyze_code or review_code
Claims → analyze_claims
Debate → analyze_debate
Logs → analyze_logs
Paper → analyze_paper
Threat report → analyze_threat_report
Product feedback → analyze_product_feedback
Sales call → analyze_sales_call

Content Creation:

PRD → create_prd
Design document → create_design_document
User story → create_user_story
Visualization → create_visualization, create_mermaid_visualization, create_markmap_visualization
Essay → write_essay
Report finding → create_report_finding
Newsletter entry → create_newsletter_entry

Improvement:

Writing → improve_writing
Academic writing → improve_academic_writing
Prompt → improve_prompt
Report finding → improve_report_finding
Code → review_code

Rating/Evaluation:

AI response → rate_ai_response
Content quality → rate_content
Value assessment → rate_value
General judgment → judge_output

2. Execute Pattern

# Basic format
fabric [input] -p [selected_pattern]

# From URL
fabric -u "URL" -p [pattern]

# From YouTube
fabric -y "YOUTUBE_URL" -p [pattern]

# From file
cat file.txt | fabric -p [pattern]

# Direct text
fabric "your text here" -p [pattern]

📚 Pattern Categories (242 Total)

Full pattern reference: read ${PAI_DIR}/skills/fabric/fabric-patterns-reference.md

Category	Count	Common Patterns
Security & Threat Modeling	15	`create_threat_model`, `create_stride_threat_model`, `analyze_incident`
Summarization	20	`summarize`, `create_5_sentence_summary`, `youtube_summary`
Extraction	30+	`extract_wisdom`, `extract_insights`, `extract_main_idea`
Analysis	35+	`analyze_claims`, `analyze_code`, `analyze_paper`
Creation	50+	`create_prd`, `create_mermaid_visualization`, `create_user_story`
Improvement	10	`improve_writing`, `improve_prompt`, `review_code`
Rating	8	`rate_content`, `judge_output`, `rate_ai_response`

🔄 Updating Patterns

The Fabric patterns exists in ~/.config/fabric/patterns/.

To see all available patterns:

ls ~/.config/fabric/patterns/

💡 Usage Examples

Threat Modeling:

# User: "Create a threat model for our new API"
fabric "API that handles user authentication and payment processing" -p create_threat_model

Summarization:

# User: "Summarize this blog post"
fabric -u "https://example.com/blog-post" -p summarize

# User: "Give me a 5-sentence summary"
fabric -u "https://example.com/article" -p create_5_sentence_summary

Wisdom Extraction:

# User: "Extract wisdom from this video"
fabric -y "https://youtube.com/watch?v=..." -p extract_wisdom

# User: "What are the main ideas?"
fabric -u "URL" -p extract_main_idea

Analysis:

# User: "Analyze this code for issues"
fabric "$(cat code.py)" -p analyze_code

# User: "Analyze these security claims"
fabric "security claims text" -p analyze_claims

🎯 Pattern Selection Decision Matrix

User Request Contains	Likely Intent	Recommended Patterns
"threat model"	Security modeling	`create_threat_model`, `create_stride_threat_model`
"summarize", "summary"	Summarization	`summarize`, `create_5_sentence_summary`
"extract wisdom", "insights"	Wisdom extraction	`extract_wisdom`, `extract_insights`
"analyze [X]"	Analysis	`analyze_[X]` (match X to pattern)
"improve", "enhance"	Improvement	`improve_writing`, `improve_prompt`
"create [visualization]"	Visualization	`create_mermaid_visualization`, `create_markmap_visualization`
"rate", "judge", "evaluate"	Rating	`rate_content`, `judge_output`
"main idea", "core message"	Core extraction	`extract_main_idea`, `extract_core_message`

🚀 Advanced Usage

Pipe content through Fabric:

cat article.txt | fabric -p extract_wisdom
pbpaste | fabric -p summarize
curl -s "https://..." | fabric -p analyze_claims

Process YouTube videos:

# Fabric handles download + transcription + processing
fabric -y "https://youtube.com/watch?v=..." -p youtube_summary

Chain patterns (manual):

# Extract then summarize
fabric -u "URL" -p extract_wisdom > wisdom.txt
cat wisdom.txt | fabric -p create_5_sentence_summary

🤖 Model Selection

Comprehensive Model Guide: read ${PAI_DIR}/skils/fabric/fabric-model-reference.md

Quick Recommendations:

High Volume/Free: gemini-2.5-flash (free, ultra-fast)
Balanced Quality: claude-sonnet-4-5-20250929 (best general purpose, recommended)
Maximum Quality: claude-opus-4-20250514 (highest quality)
Reasoning Tasks: gemini-2.0-flash-thinking-exp (extended thinking mode)
Cost Sensitive: claude-3-5-haiku-20241022 (ultra-low cost)

List Available Models: fabric --listmodels

Specify Model: fabric "content" -p pattern_name --model claude-sonnet-4-5-20250929

📖 Supplementary Resources

Strategy Reference: read ${PAI_DIR}/skils/fabric/fabric-strategies-reference.md Task-based pattern selection guide with workflows and test cases

Pattern Reference (Complete): read ${PAI_DIR}/skils/fabric/fabric-patterns-reference.md Comprehensive categorized reference for all 240 patterns

Model Reference: read ${PAI_DIR}/skils/fabric/fabric-model-reference.md Full Pattern List: ls ~/.config/fabric/patterns/ Fabric Documentation: https://github.com/danielmiessler/fabric Pattern Templates: See ~/.config/fabric/patterns/official_pattern_template/

🔑 Key Insight

The skill's value is in selecting the RIGHT pattern for the task.

When user says "Create a threat model using Fabric", your job is to:

Recognize "threat model" intent
Know available options: create_threat_model, create_stride_threat_model, create_threat_scenarios
Select the best match (usually create_threat_model unless STRIDE specified)
Execute: fabric "[content]" -p create_threat_model

Not: "Here are the patterns, pick one" Instead: "I'll use create_threat_model for this" → execute immediately

fabric

$ Installer