imagine

What problem does it solve? Image models garble exact text, numbers, and structure, and ad-hoc prompting produces inconsistent or factually wrong visuals. This Skill provides decision rules and prompt-craft guidance for the image_gen and image_edit tools so generated and edited images come out accurate, consistent, and grounded in verified facts. ## Core Features & Use Cases - Tool selection rules: Choose between image_gen for new images and image_edit for edits, iterations, and real-person depictions, with a reference-first procedure for named people. - Accuracy-first workflow: Build charts, diagrams, and text-heavy visuals with HTML/CSS code instead of image models, then verify discrete content in an inspect-and-fix loop. - Consistency and grounding: Reuse a base image across variations for recurring characters or settings, and search the web to ground factual details before prompting. - Use Case: You need a series of social media images featuring the same mascot character. Generate one base image with image_gen, then use image_edit for each variation so the character stays consistent across the set. ## Quick Start Use the imagine guidance to generate a 16:9 banner image of a sunset mountain landscape, then edit it to add a hot air balloon while keeping the composition.