Fabric

Runs any of 240+ specialized prompt patterns across extraction, summarization, analysis, creation, improvement, security, and rating. Common ones: extract_wisdom, create_threat_model, analyze_claims, improve_writing, review_code, mermaid, youtube_summary. Patterns run natively — PAI reads the pattern's system.md and applies it directly, no CLI round-trip. The fabric CLI is only used for YouTube transcripts (-y) and URL fallback (-u).

Good prompts are scattered, hard to remember, and easy to rewrite badly from scratch each time. You want a threat model, a claims analysis, or a clean summary, but reconstructing the right prompt every time is slow and inconsistent. Calling an external CLI for each one adds latency and a dependency. This skill keeps 240+ proven patterns on hand and applies them directly as prompts, so the right structured prompt is one pattern name away.

A prompt pattern system providing 240+ specialized patterns for content analysis, extraction, summarization, threat modeling, and transformation.

Patterns Location: Patterns/

Example 1: Extract wisdom from content

User: "Use fabric to extract wisdom from this article"
-> Invokes ExecutePattern workflow
-> Selects extract_wisdom pattern
-> Reads Patterns/extract_wisdom/system.md
-> Applies pattern to content
-> Returns structured IDEAS, INSIGHTS, QUOTES, etc.

Example 2: Update patterns

User: "Update fabric patterns"
-> Invokes UpdatePatterns workflow
-> Runs git pull from upstream fabric repository
-> Syncs patterns to local Patterns/ directory
-> Reports pattern count

Example 3: Create threat model

User: "Use fabric to create a threat model for this API"
-> Invokes ExecutePattern workflow
-> Selects create_threat_model pattern
-> Applies STRIDE methodology
-> Returns structured threat analysis

Pattern Execution (Native - No CLI Required)

Instead of calling fabric -p pattern_name, PAI executes patterns natively:

1. Reads Patterns/{pattern_name}/system.md
2. Applies pattern instructions directly as prompt
3. Returns results without external CLI calls

When to Use Fabric CLI Directly

Only use fabric command for:

• -y URL - YouTube transcript extraction
• -u URL - URL content fetching (when native fetch fails)

Most Common Patterns

Full Pattern Catalog

Browse the Patterns/ directory for the complete list of 240+ patterns organized by category.

How it works:

User Request → Pattern Selection → Read system.md → Apply → Return Results

Pattern Structure:

Patterns/
├── extract_wisdom/
│   └── system.md       # The prompt instructions
├── summarize/
│   └── system.md
├── create_threat_model/
│   └── system.md
└── ...240+ patterns

Each pattern's system.md contains the full prompt that defines:

• IDENTITY (who the AI should be)
• PURPOSE (what to accomplish)
• STEPS (how to process input)
• OUTPUT (structured format)

Feeds Into

• Research - Fabric patterns enhance research analysis
• Blogging - Content summarization and improvement
• Security - Threat modeling and analysis

Uses

• fabric CLI - For YouTube transcripts (-y) and URL fetching (-u)
• Native execution - Direct pattern application (preferred)

2026-01-18

• Initial skill creation (extracted from PAI/TOOLS/fabric)
• Native pattern execution (no CLI dependency for most patterns)
• Two workflows: ExecutePattern, UpdatePatterns
• 240+ patterns organized by category
• PAI Pack ready structure

• fabric -y URL for YouTube extraction — don't scrape YouTube pages. fabric handles transcript extraction natively.
• Pattern names are exact. extract_wisdom not extractwisdom. Check fabric --list if unsure.
• Long content may exceed pattern context limits. For very long inputs, chunk the content or use a summarize pattern first.