module-context

What problem does it solve?

When implementing GPU operators or fixing bugs in the Sirius GPU SQL engine, developers often lack accurate knowledge of dependency library APIs (cudf, rmm, duckdb, cucascade), leading to incorrect signatures, missed usage patterns, and version incompatibilities.

Core Features & Use Cases

• Task-to-Module Routing: Analyzes a task description against keyword mappings (joins, aggregations, sorting, memory management, pipelines, I/O) to identify which library modules are relevant.
• API Documentation Loading: Reads pre-generated per-module docs containing signatures, descriptions, and existing usage examples from the codebase.
• Structured Context Output: Produces a context block listing loaded modules, key APIs with existing usage patterns, and gap flags for unused or missing functionality.
• Use Case: When asked to implement a new hash join operator, the skill loads cudf/join, duckdb/planner, rmm/cuda_streams, and cucascade/data docs so the implementing agent follows established patterns.

Quick Start

Ask the assistant to implement a GPU aggregation operator and it will automatically load the relevant cudf, rmm, and duckdb module documentation before writing code.