phd-figure-designer

Designs and audits the three core figures of a technical research paper.

52|Updated Apr 24, 2026
One-click install
npx skills add https://github.com/Immortalqx/my_codex_skills --skill phd-figure-designer-immortalqx
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: phd-figure-designer
Source: https://github.com/Immortalqx/my_codex_skills/tree/main/phd-figure-designer
Command: npx skills add https://github.com/Immortalqx/my_codex_skills --skill phd-figure-designer-immortalqx

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Weak figures sink otherwise-strong papers: reviewers scan the Motivated Example, Solution Overview, and Experimental Results figures in under a minute to judge a submission. This Skill guides the design of those three core figures so they communicate the paper's story clearly and pass reviewer scrutiny. ## Core Features & Use Cases - Paradigm Recommendation: Selects the right design paradigm for each figure type, such as Running Example plus Failure Case for Figure 1, or Pipeline, System Architecture, and Multi-layer layouts for the methodology overview. - Layout, Labelling, and Tool Guidance: Produces a concrete layout sketch, real-entity labelling rules, font and colour-blind-safe palette guidance, and a tool recommendation (PowerPoint, Figma, draw.io, Matplotlib, TikZ). - Quality-Control Audit: Audits proposed or existing figures against universal rules covering vector format, 8pt minimum fonts, honest axes, dual encoding, and no chartjunk, with optional vision-based inspection of an uploaded image. - Use Case: Before drawing Figure 1 for a text-to-SQL paper, ask for a design and receive a three-panel failure-case layout, annotation plan, and tool suggestion, then audit the finished draft before submission. ## Quick Start Use phd-figure-designer to design Figure 1 for my paper on text-to-SQL and recommend the right paradigm, layout, and tool.

Frequently Asked Questions about phd-figure-designer

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

FAQPage Schema
How do I design Figure 1 for a research paper?

Choose one of three paradigms: Running Example plus Failure Case, Existing vs Ours, or Performance Teaser. Show a real input, highlight the current method's failure in red, and the proposed method's correct output in green, then draft in PowerPoint or Figma.

What chart type should I use for experimental results?

Match the chart to the data: grouped bars for multi-method benchmark comparisons, lines for trends and sensitivity, scatter for efficiency-versus-effectiveness trade-offs, heatmaps for matrices, and box plots for run distributions.

Matplotlib vs TikZ for paper figures, which should I use?

Use Matplotlib with a shared plot_utils.py for reproducible, scriptable experimental figures during research. Switch to TikZ or PGFPlots for camera-ready versions when exact LaTeX typography matching matters, since TikZ has a steeper learning curve.

Can the figure audit check an image I already drew?

Yes. Provide the figure image and the audit inspects font legibility, colour palette, raster-versus-vector tells, and chartjunk directly. Without an image, vision-only rules are marked as items the user must verify manually.

When should I not use this figure design skill?

Skip it for generic plotting help outside a paper, when the paper structure is not yet decided, or when reviewing a finished manuscript. Structure the paper first, then design figures, and use a pre-submission reviewer for full-paper critique.