Convert natural-language or structured requirements into deterministic, origin-anchored and auditable 2D CAD using agent-native tools. Use for drawing or validating line, circle, and arc geometry; producing DXF, AutoCAD SCR/AICAD, audit, or manifest files; fixing inaccurate AI-generated CAD; enforcing mathematical…
Build, validate, inspect, or troubleshoot deterministic SolidWorks 3D parts feature by feature from mathematical constraints. Use for AI CAD modeling requests involving SLDPRT/STEP output, extrudes and cuts, origin-anchored sketches, feature dependencies, design intent, fully constrained sketches, persistent topology…
Use before the first Image2 generation in a task to clarify intent, route to existing capabilities, compile a production-quality generation prompt, and optionally retrieve patterns from the local professional prompt library. Also use when the user asks to optimize an image prompt or choose an Image2 prompt template.
Use when the user asks AI Apprentice to design packaging, generate a packaging sample with Image2, accept visual mask corrections, locally edit the result, or hand a reviewed packaging design to AICAD for engineering drawings.
Make user-facing replies natural, direct, warm, and honest without canned service language or false human claims. Use for every conversational response, especially clarification, progress updates, corrections, failures, frustration, emotional moments, tone feedback such as "太像 AI/客服", and genuine success celebrations.