tapestry

What tapestry Does

Tapestry is a knowledge management skill that transforms scattered documents into interconnected networks of insights. Rather than treating documents as isolated files, Tapestry automatically identifies relationships between content, creates cross-references, and generates comprehensive summaries that reveal patterns and connections humans might miss. This skill is essential for product designers, researchers, and knowledge workers who manage complex information across multiple sources—whether that’s competitive analysis, user research, design systems documentation, or project archives.

The tool excels at taking your collection of documents and creating a living knowledge graph where documents inform and contextualize each other. Instead of manually tracking which document references what, Tapestry does this automatically, allowing you to query your entire knowledge base as if it were a single coherent system. This transforms how teams approach research, decision-making, and institutional knowledge management.

How to Install

1. Navigate to the Tapestry GitHub repository
2. Clone or download the tapestry-skills-for-claude-code repository to your local machine
3. Locate the tapestry folder within the downloaded repository
4. Copy the tapestry skill folder into your Claude Code skills directory (typically ~/.claude/skills/ or your configured skills path)
5. Restart Claude Code or reload your skill configuration to make Tapestry available
6. Verify installation by checking that “tapestry” appears in your available skills list
7. Begin by pointing Tapestry to a folder containing your documents (supports PDF, Markdown, plain text, and other common formats)

Use Cases

• Design System Documentation: Interlink component definitions, usage guidelines, and design tokens to create searchable documentation where changes to a base component automatically surface related impacts across your design system
• Competitive Research: Connect competitor analysis documents, market reports, and feature audits to identify patterns in competitor positioning and emerging market trends
• User Research Synthesis: Link user interviews, behavioral data, feedback surveys, and personas to discover recurring user needs and mental models across your research dataset
• Project Knowledge Base: Connect project requirements, architecture documents, technical decisions, and post-mortems so team members can understand the full context and reasoning behind any design or technical choice
• Policy and Compliance Management: Interlink regulatory requirements, internal policies, and implementation guides to ensure consistency and identify where policy changes impact multiple operational areas

How It Works

Tapestry operates in three primary phases: document ingestion, relationship mapping, and knowledge synthesis. When you provide a document collection, Tapestry first processes and indexes all content, extracting key concepts, entities, and semantic meaning. It doesn’t just look for keyword matches; it understands contextual relationships, identifying when two documents discuss the same user need even if they use different terminology.

The relationship mapping phase is where Tapestry’s intelligence shines. It creates bidirectional links between documents based on semantic similarity, explicit references, shared entities, and conceptual overlap. These connections form a knowledge graph—a visual and queryable representation of how your documents relate to each other. Unlike manual cross-referencing, this happens automatically and comprehensively across your entire collection.

Finally, Tapestry generates network summaries that explain the relationships between document clusters. When you query the system—for example, asking “What do we know about user onboarding?”—Tapestry doesn’t just return individual documents. Instead, it synthesizes insights across all connected documents, highlighting consensus, contradictions, and areas where additional research is needed. This transforms isolated documents into a unified knowledge network that’s greater than the sum of its parts.

Pros and Cons

Pros:

• Automatically discovers relationships without manual linking or taxonomy creation
• Enables powerful synthesis queries across large document collections in seconds
• Reveals blind spots and contradictions by showing all connected information at once
• Scales to hundreds of documents without degradation in usability
• Works with documents in their native formats—no conversion or data entry required
• Supports iterative refinement as you add documents and relationships evolve

Cons:

• Requires quality source documents; garbage data produces less useful networks
• Initial processing of large document collections takes time proportional to content volume
• May over-connect documents with superficial semantic similarity; manual refinement sometimes necessary
• Depends on Claude’s language capabilities; extremely specialized or niche domains may not be understood as well
• Best results require some cleanup of source documents (removing headers, metadata, formatting artifacts)

Related Skills

• Claude Search: Full-text search across documents to find specific information before Tapestry organizes it into networks
• Document Summarizer: Condense individual lengthy documents; pair with Tapestry to generate both individual and cross-document summaries
• Research Aggregator: Collect information from multiple sources; use Tapestry to interlink and organize the aggregated research
• Competitive Intelligence: Gather competitor data; Tapestry helps identify patterns and strategic implications across multiple competitive sources
• Knowledge Base Builder: Create structured documentation from your Tapestry network to share insights with your team

Alternatives

• Traditional Wikis or Confluence: Manual linking and curation required; good for published knowledge but don’t automatically discover connections in raw research or documents
• Obsidian or Roam Research: Personal knowledge management tools with manual linking; excellent for individual note-taking but require deliberate structure creation and don’t automatically extract relationships from bulk document imports