v1.1.1

Changed

• make-a-deck: clarify that text-capable image generation is not a reason to prefer it — HTML/CSS stays the default for structured and precise slide content, and generated imagery remains a supplement reserved for conceptual scenes, mascots, hero/section art, and genuine infographics (Chinese labels are welcome when an infographic is genuinely the right call).

v1.1.0

Changed

• generate-images: update raster image guidance for modern image backends that can render text reliably, including Chinese / CJK, and route text-heavy infographics to HTML/CSS only when editability, exactness, or data binding matters.
• make-a-deck: align deck imagery guidance with text-capable image generation so generated slide infographics can include headings, labels, callouts, and Chinese copy when they do not need live editability or exact numeric binding.

v1.0.0

1.0.0 - 2026-06-22

Features

• Ship the portable baoyu-design Agent Skill with the core design methodology, harness references for Claude Code, Cursor, and Codex, built-in design workflows, starter components, design-system tooling, and local import/export agents.
• Add release-skills, a reusable release workflow for version selection, multi-language changelogs, annotated tags, and GitHub Releases.
• Add local deck, document, PPTX, video, image-generation, Figma .fig, GitHub repo, HTML/CSS import, and design-system preview flows.

Fixes

• gen-pptx: improve editable PowerPoint fidelity by rendering children in z-index order, propagating ancestor opacity through exported subtrees, rasterizing gradient overlays, preserving Unicode filenames, preserving line breaks, and guarding stuck captures with per-call timeouts.
• make-a-deck: document full-height slide wrappers so full-bleed slide content exports reliably.

Documentation

• Document installation, local usage, design systems, import sources, deck/PPTX export, and the supported agent harnesses in the English and Chinese READMEs.