Ask Codex to turn your research notes or paper into a branded SJTU presentation deck automatically.
Generate multiple presentation variations from the same content while keeping the university's visual identity.
Enhance data-heavy slides with Nature-style scientific charts using the bundled workflow.
Maintain hand-edited slides safely without the AI agent overwriting your manual changes.
Requires OpenAI Codex access; user pastes a ready-made install prompt into Codex to fetch and register the skill from GitHub.
sjtu-ppt-template-skill is a Codex skill for producing editable PowerPoint decks in the visual style of Shanghai Jiao Tong University, a major research university in China. Codex is OpenAI's coding agent, and skills are bundles of instructions, reference notes, and assets that the agent loads on demand. Once installed, the user can ask Codex to make an SJTU-style presentation from their notes, slides, or paper, and the agent follows the playbook in this repository instead of trying to design something from scratch. The skill bundles ready-to-use materials under assets/. That includes SJTU logo PNG files, PPTX templates, and an optional font package, so the agent does not have to find or guess at the brand visuals. The README points out that this is a personal project rather than an official university release, and asks anyone using it publicly, commercially, or for redistribution to confirm asset permissions first. A separate ASSET_NOTICE.md file spells out the terms. Beyond the visual templates, the skill packages a set of reference workflows. There is a non-destructive revision policy under references/revision-safety.md, meant to stop the agent from overwriting edits the user already made by hand. There are notes on choosing the right template for the topic, on writing speaker notes and a presentation script, on quality gates the deck should pass before being handed back, and on optional Nature-style scientific chart enhancement for slides that include data figures. The README provides a ready-made install prompt, in both English and Chinese, that the user pastes into Codex to fetch the skill from GitHub and register it locally.
Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.