What problem does it solve? Processing many independent documents or repository questions serially wastes time when the work could be split between the active primary and a delegate worker. This Skill coordinates that split so both sides work in parallel with clear attribution of who produced what. ## Core Features & Use Cases - Parallel fan-out: Split a supplied corpus of documents or a sealed repository into independent work units, dispatching one portion to an eligible delegate worker while the active primary processes the rest simultaneously. - Bounded source binding: Each work unit carries its own bounded UTF-8 documents or a canonical repository root with an exact expected head, so the delegate never browses for missing sources. - Attributed synthesis: Merge both halves into one response with explicit annotation of which items came from the delegate, preserving the user's per-model calibration and audit trail. - Use Case: Given six competitor reports, send three to the delegate for product, audience, and price extraction while the active primary handles the other three, then merge the tables with per-item attribution. ## Quick Start Ask the assistant to delegate half of your supplied documents to the reviewer for parallel extraction and merge the results with attribution.
