What problem does it solve? Processing hundreds or thousands of documents, records, or rows one at a time is slow and error-prone. This Skill splits massive independent data tasks into batches, deploys parallel sub-agents to process them, and aggregates validated results with failure recovery. ## Core Features & Use Cases - Swarm Design & Sizing: Computes batch size from token budgets and swarm size from total items, capping at 20 agents per wave with multi-wave deployment for larger jobs. - Structured Agent Briefs: Generates self-contained briefs per agent with input data, output schemas, quality rules, and error protocols so results merge reliably. - Aggregation & Recovery: Validates schema compliance, completeness, and duplicates, then retries failed items up to two times before marking them unrecoverable. - Use Case: Score 2,000 sales leads against ICP criteria by splitting them into batches of 40, deploying 50 agents across 3 waves, and merging the scored results into a single CSV with a failure report. ## Quick Start Use the agent-swarm-deployer skill to analyze the sentiment of all 1,000 customer reviews in reviews.csv and output the results as a single CSV.
