stevenjinlong

Create a PowerPoint deck by first generating a standard image-first PPTX, then running a ppt-master reconstruction pass to produce a native editable PPTX. Use when the user invokes $awesome-ppt-editable or explicitly wants gpt-image/imagegen slide images converted into an editable PPT with native text, shapes, charts, and objects. This is a reconstruction workflow, not automatic lossless raster-to-vector conversion.

Core Awesome PPT workflow for image-model-rendered PowerPoint decks. Use when the user invokes $awesome-ppt or asks to make a polished PPT/PPTX by designing full-slide image prompts where gpt-image/imagegen renders the complete slide including all titles, body text, labels, and visual design, assembling those rendered slide images into a PPTX, and, only in editable mode, using ppt-master to rebuild the image-rendered deck as native editable text, shapes, charts, and objects.

Create a standard image-first PowerPoint deck from a prompt, brief, or source files. Use when the user invokes $awesome-ppt-std or wants a polished PPT/PPTX where gpt-image/imagegen renders complete full-slide images including all visible text, labels, diagrams, and visual design, then assembles those rendered images into a PPTX. This mode does not run ppt-master and does not claim fully native editability.

Generate or edit raster images through the OpenAI-compatible image endpoint configured in the user's local Codex installation. Use when the current Codex environment relies on a custom or self-hosted `base_url`, not the official OpenAI endpoint, and image generation should go through that remote provider rather than the built-in image pipeline. This skill is especially relevant for text-to-image posters, illustrations, sprites, mockups, product shots, and reference-image edits sent through the configured remote endpoint.