Knowledge Sync Strategy

Knowledge Sync Strategy Workflow

This Workflow provides Claude Code with step-by-step guidance for implementing Knowledge workspace sync.

When to use this workflow

• When implementing Knowledge workspace file system sync
• When integrating RAG pipeline with file synchronization
• When handling source document imports (PDF, URL, local files)
• When managing embeddings sync with document changes
• When creating sync status indicators for Knowledge workspace

Workflow Steps

Step 1: Knowledge File System Strategy (2-3 hours)

Agent: file-sync-specialist Input: Knowledge workspace requirements Output: File system strategy document

Acceptance Criteria:

• Desktop: FSA API with source document storage
• Mobile: IndexedDB with sync-to-FSA on desktop
• Document format support (PDF, Markdown, plain text)
• Folder structure defined (collections, sources, embeddings)
• Permission handling (read/write per collection)

Step 2: RAG Pipeline Integration (4-6 hours)

Agent: workspace-architect Input: File system strategy Output: RAG sync implementation

Acceptance Criteria:

• Document indexing (chunking, embedding, storage)
• Incremental updates (only reindex changed documents)
• Embedding sync (Orama WASM with IndexedDB)
• Search integration (RAG query with file system)
• Source import handling (PDF parser, URL fetcher, local file reader)

Step 3: AI Synthesis Integration (4-6 hours)

Agent: workspace-architect Input: RAG sync implementation Output: AI integration with Knowledge workspace

Acceptance Criteria:

• Per-document AI synthesis (summarize, extract key points)
• Batch AI synthesis (process entire collection)
• RAG-powered search (semantic search with citations)
• AI tool permissions (read documents, write summaries)
• Progress indicators for long-running operations

Step 4: Sync Status UX/UI (2-3 hours)

Agent: component-splitter Input: AI integration Output: Sync status UI components

Acceptance Criteria:

• Sync status indicator (syncing, synced, indexing, offline)
• Import progress UI (source import with progress bar)
• Indexing progress UI (chunking, embedding, indexing stages)
• Search results UI (semantic search with citations)
• Error states (import failed, indexing failed, retry button)

Step 5: Validation & Testing (2-3 hours)

Agent: test-writer Input: Complete Knowledge sync implementation Output: Test suite + validation report

Acceptance Criteria:

• Unit tests for sync logic (CRUD, RAG integration)
• Integration tests (sync + embeddings + search)
• E2E tests (critical user journeys: import, search, AI synthesis)
• Test coverage ≥80%
• Zero data loss scenarios validated

Example Usage

"Implement Knowledge sync strategy using the knowledge-sync-strategy workflow"

Validation Commands

# TypeScript check
pnpm tsc --noEmit --incremental

# Test suite
pnpm test

# Knowledge-specific tests
pnpm test knowledge-*

# RAG validation
pnpm test rag-*

# Sync validation
pnpm test sync-knowledge-*

Success Criteria

• ✅ Knowledge file system sync working (read, write, sync)
• ✅ RAG pipeline integrated (indexing, search, retrieval)
• ✅ Source imports working (PDF, URL, local files)
• ✅ AI integration working (per-document, batch, RAG search)
• ✅ Sync status indicators visible (syncing, indexing, offline)
• ✅ Embeddings sync working (Orama WASM with IndexedDB)
• ✅ Zero data loss scenarios
• ✅ Test coverage ≥80%
• ✅ E2E tests passing

Artifacts

For detailed workflow documentation, refer to: Knowledge Sync Strategy Workflow

Converted and distributed by TomeVault — claim your Tome and manage your conversions.