Data Throughput Accelerator

When you're staring at a multi-hour backfill or a warehouse load that should take minutes but doesn't, this skill separates the actual bottlenecks (source extraction, network, transforms, warehouse writes) and applies the right fix to each. It pushes you toward warehouse-native operations, manifest checkpoints, idempotent writes, and proper partitioning instead of throwing more workers at the problem. The workflow is measure backlog, benchmark variants, promote the fastest correct path, then codify it as a runbook or scheduled job. The accounting block format is strict: source files discovered, rows added, remaining tail, runtime, and a correctness gate that confirms manifests and table timestamps actually match. No hiding failed files or deleting raw data to game the metrics.

Install to Claude Code

npx -y skills add affaan-m/everything-claude-code --skill data-throughput-accelerator --agent claude-code