paper-visualization
Skill9 starsChanged 3 months agoMITView on GitHub
Route paper figure work to the right lightweight path and enforce manuscript-grade visual rules. Use when user says \"论文作图\", \"补图\", \"论文可视化\", \"画论文图\", \"paper visualization\", \"figure pipeline\", or wants one entry point for plots, architecture figures, and figure audits.
In this repository47 other files it gives its agents.
- autodl-hpcSkill
- experiment-bridgeSkill
- experiment-planSkill
- human-toneSkill
- idea-discoverySkill
- paper-writingSkill
- rebuttalSkill
- research-blueprintSkill
- research-pipelineSkill
- research-repo-architectSkill
- research-reviewSkill
- resubmit-pipelineSkill
- autodl-hpcSkill
- auto-paper-improvement-loopSkill
- auto-review-loopSkill
- novelty-checkSkill
- paper-figureSkill
- paper-planSkill
- paper-writeSkill
- research-refineSkill
- research-reviewSkill
- experiment-bridgeSkill
- experiment-planSkill
- auto-paper-improvement-loopSkill
- auto-review-loopSkill
- grant-proposalSkill
- idea-creatorSkill
- idea-discoverySkill
- novelty-checkSkill
- paper-figureSkill
- paper-planSkill
- paper-posterSkill
- paper-slidesSkill
- paper-writeSkill
- paper-writingSkill
- research-refineSkill
- research-reviewSkill
- human-toneSkill
- idea-discoverySkill
- paper-visualizationSkill
- paper-writingSkill
- rebuttalSkill
- research-blueprintSkill
- research-pipelineSkill
- research-repo-architectSkill
- research-reviewSkill
- resubmit-pipelineSkill
The file
We split this skill into the rules it gives an agent
9 rules. About 0 are common, 4 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 34
Vector first: prefer PDF / SVG / native LaTeX vector output over raster.
- Very specific: line 38
Local-lightweight default: do not invent heavyweight GUI workflows when paper-figure, figure-spec, or template-backed LaTeX/matplotlib can …
- Very specific: line 66
Use /paper-illustration or /paper-illustration-image2 only when the figure is not naturally expressed as data plots or deterministic …
- Very specific: line 70
Use /mermaid-diagram only for low-stakes drafts, planning, or quick visual communication.
- Very specific: line 74
After generation, run /figure-table-audit on the final figure set.
- Distinctive: line 78
When a user asks for "paper figures" without enough detail:
- Very specific: line 80
Read PAPER_PLAN.md, paper/, figures/, and any existing figure plan.
- Distinctive: line 118
Do not stop at "use skill X".
- Distinctive: line 120
if the user asked only for process guidance, return the routed figure plan and the typography/export constraints.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Paper Visualization
Capability Routing
This is a first-layer entry skill. Keep it loaded as the user-facing route; when a request needs a specialized capability below, resolve the debuffer repo root from .debuffer_skills/installed-skills-codex.txt (repo_root) when available, read the referenced library SKILL.md, then follow that skill. Do not copy the whole library skill into this file.
/paper-figure: read../library/paper/paper-figure/SKILL.md./figure-spec: read../library/paper/figure-spec/SKILL.md./image-to-vector-ppt: read../library/paper/image-to-vector-ppt/SKILL.md./figure-table-audit: read../library/paper/figure-table-audit/SKILL.md./mermaid-diagram: read../library/paper/mermaid-diagram/SKILL.md./render-html: read../library/paper/render-html/SKILL.md./paper-illustration: read../library/paper/paper-illustration/SKILL.md./paper-figure-artifact-audit: read../library/paper/paper-figure-artifact-audit/SKILL.md.
Route paper-figure work for: $ARGUMENTS
This is a lightweight orchestration skill. It does not replace the existing renderers. It decides which skill should own each figure and applies one shared publication rule set before any drawing starts.
Shared Rules
- Body-font lock: any visible text inside a figure must match the paper body font size unless the venue explicitly requires a documented exception.
- Vector first: prefer PDF / SVG / native LaTeX vector output over raster.
- No title inside figure: titles belong in captions, not inside the image.
- Reproducible artifacts: save the generating script/spec next to the output.
- Local-lightweight default: do not invent heavyweight GUI workflows when
paper-figure,figure-spec, or template-backed LaTeX/matplotlib can do the job.
Default Routing
Choose the narrowest route that matches the requested figure:
- Data plots and result figures
Use/paper-figure. Typical cases: line/bar/scatter/heatmap/box plots, ablations, scaling curves, comparison tables, multi-panel result figures. - Architecture / workflow / system / pipeline diagrams
Use/figure-specfirst. Choose this when layout is structured and should stay deterministic, editable, and easy to revise. - Native LaTeX vector figures
Stay in the project with PGFPlots/TikZ when:- the figure should inherit manuscript typography exactly,
- the figure is best expressed as axis-based LaTeX graphics,
- or the target venue is sensitive to font consistency. Prefer:
templates/figure/matplotlib_publication_style.pytemplates/figure/pgfplots_bodyfont.tex
- Qualitative concept illustrations
Use/paper-illustrationor/paper-illustration-image2only when the figure is not naturally expressed as data plots or deterministic diagrams. - Simple temporary flowcharts
Use/mermaid-diagramonly for low-stakes drafts, planning, or quick visual communication. Do not treat Mermaid as the default for submission figures. - Submission-quality check
After generation, run/figure-table-auditon the final figure set.
Triage Procedure
When a user asks for "paper figures" without enough detail:
- Read
PAPER_PLAN.md,paper/,figures/, and any existing figure plan. - Build a compact figure inventory:
- figure id
- purpose / claim supported
- preferred route
- status: existing / generate / revise / audit
- Generate only the missing or weak figures first.
- Audit before handing the figure set to
/paper-writeor/paper-writing.
Practical Defaults
- For plots, prefer the style habits adapted from
MLNLP-World/Paper-Picture-Writing-Code, but normalized throughtemplates/figure/matplotlib_publication_style.py. - For PGFPlots/TikZ, import
templates/figure/pgfplots_bodyfont.texand keep the figure text size equal to manuscript body text. - For architecture figures, prefer
/figure-specover ad hoc draw.io-style manual work unless the user explicitly asks for a manual editing workflow.
Output
Produce a compact decision record in the working notes or response:
Figure inventory:
- Fig 1: architecture overview -> /figure-spec -> generate
- Fig 2: main results bar chart -> /paper-figure -> generate
- Fig 3: ablation heatmap -> /paper-figure -> revise
- Table 1: benchmark comparison -> /paper-figure -> generate
Global rules:
- figure text size = paper body text size
- vector-first export
- final pass: /figure-table-auditCompletion Rule
Do not stop at "use skill X". Either:
- invoke the downstream skill immediately, or
- if the user asked only for process guidance, return the routed figure plan and the typography/export constraints.
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
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.

