Add pgvector and hybrid text search skills

[cevian]

Summary

• Add pgvector-semantic-search skill: comprehensive guide for pgvector setup, HNSW indexes, quantization strategies, filtering, and troubleshooting
• Add hybrid-text-search skill: combining pg_textsearch (BM25) with pgvector using RRF fusion, with Python and TypeScript examples

Skills Added

pgvector-semantic-search (323 lines)

• Golden path defaults (halfvec, HNSW, cosine distance)
• HNSW and IVFFlat index configuration
• Binary quantization for large datasets
• Filtering best practices (iterative scan, partial indexes, partitioning)
• Monitoring, debugging, and common issues

hybrid-text-search (241 lines)

• When to use hybrid vs semantic-only vs keyword-only
• Parallel query pattern with client-side RRF fusion
• Weighting and reranking with cross-encoders
• Python and TypeScript code examples
• BM25 configuration notes (k1/b tuning, partitioned tables)

🤖 Generated with Claude Code