Pinecone

This gives Claude Code the ability to work with Pinecone's managed vector database service. It's aimed at production RAG applications where you need auto-scaling and low latency without managing infrastructure yourself. The skill supports hybrid search with both dense and sparse vectors, which is useful if you're combining semantic and keyword search. Pinecone promises sub-100ms p95 latency and handles billions of vectors, so this makes sense if you're beyond the prototyping stage and need something that won't fall over under load. The skill template comes from davila7's collection but originated from ovachiever's droid-tings repo. Worth noting it failed Snyk's security audit while passing the others.