This is the tool you want when a ClickHouse query is dragging or you need to find which queries are eating your cluster alive. It follows an evidence-first workflow: you feed it a slow query or query_id, it pulls execution stats and table schema details, then suggests concrete fixes ranked by impact and risk. The pre-flight logic is smart enough to ask what you're optimizing for (duration, CPU, memory, disk) if you don't have a specific query in hand. It defaults to a light evidence pass to keep things fast, only going deep when you need raw details. All recommendations get validated before it hands them back, and it knows the difference between minmax, bloom_filter, and set indexes based on your actual predicates.
npx -y skills add frankchen021/datastoria --skill optimize-clickhouse-sql --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
JamieMason/syncpack
github/awesome-copilot
addyosmani/agent-skills