Embeddings
Initialize claude-flow embeddings, index text with HNSW, and run semantic search or memory-backed similarity for agent knowledge.
Overview
embeddings is an agent skill most often used in Build agent-tooling (also Operate iterate) that sets up claude-flow vector embeddings, HNSW search, and semantic memory for agents.
Install
npx skills add https://github.com/ruvnet/ruflo --skill embeddingsWhat is this skill?
- sql.js persistent cache with HNSW indexing for fast similarity search
- CLI: init, embed, batch, and search with top-k queries
- Memory store/search integration with optional --embed and --semantic flags
- Normalization, chunking, quantization, and hyperbolic embedding options
- Documented 75x speed claim with agentic-flow ONNX integration
- 75x faster cited with agentic-flow ONNX integration
- HNSW described as 150x–12,500x faster search
- Feature table covers sql.js, hyperbolic, normalization, and chunking
Adoption & trust: 680 installs on skills.sh; 58.5k GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
Your agent cannot find related patterns, docs, or prior decisions because keyword search misses meaning across large knowledge stores.
Who is it for?
Indie agent builders adding semantic search, pattern matching, or RAG-style memory to claude-flow or ruflo projects.
Skip if: Simple exact-match lookups, tiny repos with no corpus, or teams unwilling to run npx claude-flow embedding CLI steps.
When should I use this skill?
Semantic search, pattern matching, similarity queries, or knowledge retrieval; skip for exact text matching only.
What do I get? / Deliverables
You get initialized embedding storage, batch-indexed content, and top-k semantic queries wired into claude-flow memory for agent retrieval.
- Initialized embeddings backend (sqlite)
- Embedded documents or memory keys with vectors
- Top-k semantic search results for agent queries
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Primary shelf is Build agent-tooling because the skill configures vector search infrastructure agents rely on while constructing products. Agent-tooling covers embedding pipelines, persistent caches, and semantic retrieval hooks for coding agents.
Where it fits
Initialize sqlite embeddings and batch-index SKILL.md corpora before shipping an internal coding agent.
Semantic-search stored runbooks when debugging recurring production issues.
Embed competitor notes and query top-k similar positioning angles during discovery.
Retrieve semantically related FAQ entries when drafting support replies.
How it compares
Local claude-flow embedding stack with HNSW—not a hosted-only vector SaaS or pure SQL full-text search.
Common Questions / FAQ
Who is embeddings for?
Solo builders and agent authors using ruflo/claude-flow who need semantic search, similarity, and embedded memory without a separate vector DB service.
When should I use embeddings?
Use it while building agent tooling to index docs and patterns, during operate iteration to query institutional knowledge, or whenever SKILL.md triggers like semantic search and pattern matching apply.
Is embeddings safe to install?
It runs local CLI and sqlite WASM cache commands; review the Security Audits panel on this page and scope what text you embed if it contains secrets.
SKILL.md
READMESKILL.md - Embeddings
# Embeddings Skill ## Purpose Vector embeddings for semantic search and pattern matching with HNSW indexing. ## Features | Feature | Description | |---------|-------------| | **sql.js** | Cross-platform SQLite persistent cache (WASM) | | **HNSW** | 150x-12,500x faster search | | **Hyperbolic** | Poincare ball model for hierarchical data | | **Normalization** | L2, L1, min-max, z-score | | **Chunking** | Configurable overlap and size | | **75x faster** | With agentic-flow ONNX integration | ## Commands ### Initialize Embeddings ```bash npx claude-flow embeddings init --backend sqlite ``` ### Embed Text ```bash npx claude-flow embeddings embed --text "authentication patterns" ``` ### Batch Embed ```bash npx claude-flow embeddings batch --file documents.json ``` ### Semantic Search ```bash npx claude-flow embeddings search --query "security best practices" --top-k 5 ``` ## Memory Integration ```bash # Store with embeddings npx claude-flow memory store --key "pattern-1" --value "description" --embed # Search with embeddings npx claude-flow memory search --query "related patterns" --semantic ``` ## Quantization | Type | Memory Reduction | Speed | |------|-----------------|-------| | Int8 | 3.92x | Fast | | Int4 | 7.84x | Faster | | Binary | 32x | Fastest | ## Best Practices 1. Use HNSW for large pattern databases 2. Enable quantization for memory efficiency 3. Use hyperbolic for hierarchical relationships 4. Normalize embeddings for consistency