paper-talk

Generates conference talk slides, speaker notes, and audit reports from a compiled academic paper.

Updated Jul 9, 2026
One-click install
npx skills add https://github.com/Lingjie-wang/autoRL --skill paper-talk-lingjie-wang
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: paper-talk
Source: https://github.com/Lingjie-wang/autoRL/tree/main/Auto-claude-code-research-in-sleep/skills/paper-talk
Command: npx skills add https://github.com/Lingjie-wang/autoRL --skill paper-talk-lingjie-wang

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python-pptx.

What problem does it solve? Turning a finished academic paper into a conference-ready talk requires manually building slide outlines, Beamer and PPTX decks, speaker notes, and verifying that every claim, citation, and anonymity placeholder on the slides is correct. This Skill automates that entire pipeline from paper directory to final exported talk artifact. ## Core Features & Use Cases - End-to-end talk pipeline: Runs six phases from slide outline generation through Beamer/PPTX build, per-page visual polish, assurance audits, export integrity checks, and a final report. - Assurance ladder: Choose draft, polished, or conference-ready levels; the highest level adds claim audits, citation verification via DBLP/CrossRef, and anonymity scans on slide text, speaker notes, and talk script. - Content integrity guarantees: The original paper and baseline deck are never modified, speaker notes are byte-stable across polish, and no new content is invented anywhere in the pipeline. - Use Case: You have a compiled NeurIPS paper in paper/ and need a 15-minute spotlight talk. Run the pipeline with assurance: conference-ready to get a polished PPTX, full talk script, and audit JSONs confirming every slide claim is paper-grounded. ## Quick Start Ask the AI to run the paper-talk pipeline on your compiled paper directory with talk type spotlight, 15 minutes, and conference-ready assurance to produce the full slide deck, speaker notes, and audit report.

Frequently Asked Questions about paper-talk

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate conference talk slides from my paper?

Run the pipeline against your compiled paper directory containing main.tex and main.pdf. It produces a slide outline for approval, then builds Beamer source, an editable PPTX, speaker notes, and a full talk script under the slides/ directory.

What is the difference between draft, polished, and conference-ready assurance levels?

Draft builds outline, deck, and export only. Polished adds per-page visual polish against a reference PDF. Conference-ready additionally runs claim audits, citation verification, and anonymity scans on slide text, notes, and script before granting the final verdict.

Does the slide generation modify my original paper files?

No, the paper directory is strictly read-only throughout the pipeline. The workflow consumes main.tex, sections, and figures but never modifies them, and the baseline deck is preserved with polish writing a separate versioned copy.

What tools are required to run the conference talk pipeline?

You need a LaTeX toolchain (xelatex, pdflatex, or latexmk), LibreOffice headless for PPTX-to-PDF rendering, poppler utilities (pdfinfo, pdftoppm) for PNG previews, python-pptx, and Codex MCP access for per-page polish calls.

When should I not use this talk pipeline?

Skip it when the paper is not yet compiled, when you need a poster instead of a talk, when you already have a finished deck needing only visual polish, or when the talk content is unrelated to a paper such as a general lecture.

How does the anonymity check work for anonymous submissions?

The anonymity scan examines slide text, speaker notes, and talk script for real paper titles, author names, submission counts, or deanonymizing URLs where placeholders should appear. Any failure blocks the conference-ready verdict and halts for human review.