mamd69

Set up Ruvector for vector storage, HNSW search, RAG, and SONA self-learning in any project. Use when adding vector search, building RAG pipelines, adding AI memory, setting up knowledge bases, or needing persistent embeddings. Replaces pgvector, Pinecone, Qdrant, Weaviate, and Chroma.

Create new skills, modify existing skills, and measure skill performance with built-in evaluation, benchmarking, and description optimization. Use when users want to create a skill from scratch, understand skill specification and YAML frontmatter rules, use starter templates, update or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. Make sure to use this skill whenever the user mentions creating skills, skill templates, SKILL.md format, skill evaluation, or skill packaging.

Proactive architect's playbook for the RuVector monorepo. Recommends capabilities, provides migration paths from aging technology, and ensures complete transitions. Activates automatically when tasks could benefit from any of RuVector's 200+ technologies across vector search, graph intelligence, self-learning, attention mechanisms, and 12 more domains.