ppt-expert
Run the following command in DeepSeek Harness:
dsh plugin install TANGZHUO12/ppt-expert
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install TANGZHUO12/ppt-expert to install; source code is at https://github.com/TANGZHUO12/ppt-expert — after install, execute setup.sh to provision the MCP server.
About this plugin
Having an AI build slides usually falls short because it lacks a real design sense. ppt-expert wires a 247-line slide-expert persona, nine LibreOffice Impress MCP tools, a matplotlib chart kernel, and a live browser preview into a single dsh capability bundle — one prompt in, and the agent lays out pages, adds animations, and renders business charts page by page, with the preview flipping forward every time a page is saved.
The whole chain carries zero npm dependencies. Charts are produced by matplotlib (18 business types, 40-120 ms each, native CJK), tables are read from xlsx / csv / ods, and the slide grid is calibrated to a 25.4×14.29 cm standard with 11 hard rules that include automated contrast auditing — so the deck holds up on a projection screen, not just on your laptop.
Built for dsh workflows where decks are a recurring deliverable: pre-sales proposals, product launches, quarterly reports, and training materials. One sentence in, watch it assemble in real time, skip the template-hunting loop entirely.
Use Cases
- Generate a full animated deck with business charts from a single prompt
- Render matplotlib charts from xlsx or csv data and embed them into slides
- Watch the agent lay out each page with a live browser preview that auto-flips
- Calibrate to a 25.4x14.29 cm grid for reliable on-stage projection
Best For
- Teams producing business decks regularly inside dsh workflows
- Individuals drafting pre-sales decks, quarterly reports, or product launches with AI
- Slide designers who want to skip the template-hunting and color-wrangling loop
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.