agentleFS
wangyi0403wangyi0403/dl-research-workflow/.agents/skills/paper-illustration/SKILL.md

paper-illustration

Skill4 starsChanged 41 days agoMITView on GitHub

Create non-data scientific figures such as method diagrams, mechanisms, and conceptual schematics; use editable vector tools when appropriate.

6 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create or guide the creation of non-data scientific figures while preserving an editable source, claim linkage, and visual verification.

    line 8
  2. Distinctive:

    Every figure must identify its Claim/Chain ID, narrative role, source material, and limits of interpretation.

    line 15
  3. Very specific:

    Update docs/10_figure_report.md with the Claim/Chain ID, source path, preview/export path, caption status, QA result, and remaining author …

    line 56
  4. Distinctive:

    Before final production, verify the figure contract addresses:

    line 62

“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Paper Illustration

Create or guide the creation of non-data scientific figures while preserving an editable source, claim linkage, and visual verification. A prompt is an intermediate artifact, not the default final deliverable.

Principles

  • Choose colors, typography, and encodings from the project style and target venue; do not impose universal role colors or a universal font family.
  • Preserve consistent terminology, visual hierarchy, line weights, and accessible redundant encodings across the manuscript.
  • Choose editable vector diagrams when precise structure or future editing matters. For non-data concept, method or workflow figures, honor an explicit user preference for built-in image generation; verify every label and relationship against the figure contract. Generated pixels are not an editable vector source.
  • Every figure must identify its Claim/Chain ID, narrative role, source material, and limits of interpretation.

Process

For each figure in the paper plan:

1. Freeze the figure contract

Record:

  • Figure content and layout
  • project/venue style constraints
  • Spatial arrangement
  • What each element represents

Save the contract or generation notes to results/figures/0X_xxx.md when a persistent artifact is needed.

Done when: the figure purpose, claim, content, layout, terminology, style constraints, and acceptance criteria are explicit.

2. Select the production route

  • Architecture, flow, process, concept, or editable schematic: choose a diagram capability when editability is required; an explicitly requested image-generation route may supply the visual version. If both versions are retained, keep their semantics consistent.
  • Composed product/UI view: use the available design capability when appropriate.
  • Scientific illustration or bitmap material: use the available image-generation capability when it matches the content or the user's requested visual style. Quantitative charts remain generated from verified data and code.
  • User hand-drawing: use a detailed prompt/reference pack only when the user explicitly chooses that route.

Done when: the selected route matches the scientific content and required editability.

3. Produce the figure

Create the requested final asset and a reviewable preview using the selected capability. Retain an editable source when required; do not describe a generated bitmap as editable. For a cropped draw.io figure intended for LaTeX, use the installed CLI's --crop PDF export when supported; inspect the complete diagram and its PDF page bounds before inclusion, because a default page export may clip wide diagrams. Keep generated assets, prompts, and sources under results/figures/ with stable descriptive names. Do not fabricate scientific objects, causal links, quantitative values, or experimental outcomes for visual completeness.

Done when: the requested asset and preview exist, any required editable source is available, or the user has explicitly requested a prompt-only handoff.

4. Verify

Inspect the rendered figure at expected manuscript size. Check clipping, overlap, reading order, terminology, arrow meaning, color accessibility, caption consistency, and whether the visual supports no stronger claim than the evidence permits.

Done when: programmatic or visual QA is complete and material uncertainties are recorded.

5. Record and hand off

Update docs/10_figure_report.md with the Claim/Chain ID, source path, preview/export path, caption status, QA result, and remaining author decisions.

Done when: the final figure and editable source are traceable and pass the Figure Design Checklist.

Figure Design Checklist

Before final production, verify the figure contract addresses:

  • Clear visual hierarchy (most important element largest/most prominent)
  • Color semantics consistent with paper conventions
  • All text readable at expected print size and compliant with the target venue
  • No unnecessary decorative elements
  • Self-contained: understandable without reading the paper
  • Journal-compliant: no color-only information (colorblind-safe)

Output

  • results/figures/0X_xxx.md — figure contract or prompt/reference pack when needed
  • editable vector/design source for structured figures
  • reviewable PNG/PDF/SVG export as appropriate
  • optional results/figures/0X_material_* assets with provenance
  • corresponding entry in docs/10_figure_report.md

References

FileOpen when
references/figure-type-selection.mdDeciding which figure type fits the content.
references/architecture-patterns.mdChoosing visual layout. 7 common scientific figure motifs (pipeline, comparison, zoom-in, etc.) with rules and anti-patterns.
references/color-palette.mdSelecting a project-appropriate accessible palette and checking grayscale interpretation.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.