weave-cli-skills

Evaluate and benchmark RAG solutions using weave-cli. Use when: (1) creating evaluation datasets, (2) running evals against agents, (3) benchmarking multiple agents, (4) creating custom evaluators (LLM judge or human), (5) analyzing eval results and recommending improvements. NOT for: initial setup (use weave-setup), data ingestion (use weave-ingest), querying (use weave-search), or deployment (use weave-stack).

Ingest data into vector databases using weave-cli. Use when: (1) creating collections and schemas, (2) ingesting documents (PDF, text, images), (3) batch pipeline ingestion with workers, (4) getting AI schema/chunking suggestions, (5) backing up and restoring collections. NOT for: initial setup (use weave-setup), querying data (use weave-search), evaluations (use weave-eval), or deployment (use weave-stack).

Plan and orchestrate end-to-end RAG solutions using weave-cli. Use when: (1) gathering user requirements for a RAG project, (2) designing the solution architecture (VDB, embeddings, agents, evals), (3) creating execution plans across multiple agents, (4) coordinating the RAG lifecycle from setup through production, (5) recommending improvements after eval results. This is the orchestrator skill — it coordinates work across weave-setup, weave-ingest, weave-search, weave-eval, and weave-stack.

Query vector databases and manage AI agents using weave-cli. Use when: (1) running semantic, BM25, or hybrid searches, (2) creating and configuring RAG/QA/summarize agents, (3) comparing embedding models, (4) re-embedding collections, (5) tuning search parameters. NOT for: initial setup (use weave-setup), data ingestion (use weave-ingest), evaluations (use weave-eval), or deployment (use weave-stack).

Set up and configure weave-cli for vector database RAG solutions. Use when: (1) installing or updating weave-cli, (2) creating or updating config.yaml and .env files, (3) running weave doctor diagnostics, (4) selecting and configuring vector databases, (5) verifying health of VDB connections. NOT for: data ingestion (use weave-ingest), querying (use weave-search), evaluations (use weave-eval), or stack deployment (use weave-stack).

Deploy and operate RAG infrastructure using weave-cli stacks. Use when: (1) initializing weave-stack.yaml for k8s or podman, (2) deploying VDB + ingestion + dashboard, (3) monitoring stack health, (4) day-2 operations (backup, scaling, security), (5) managing the weave dashboard UI. NOT for: initial setup (use weave-setup), data ingestion (use weave-ingest), querying (use weave-search), or evaluations (use weave-eval).