- Home
- Skills
- Research & Analysis
- Figure Editor: Multi-Panel Journal Figures from Your Images
More screenshots
Works with the AI Agents you already use
Figure Editor: Multi-Panel Journal Figures from Your Images
Turn the plots, micrographs and diagrams you already have into one labelled multi-panel figure at exact journal width, with placeholders for missing panels, a print-resolution check and a caption draft.
$9
Figure Editor: Multi-Panel Journal Figures from Your Images
Example session with this skill installed
Put Figure 3 together for my journal submission: double-column (two-column) width, bold lowercase labels. Panels are in research/figures/, in this order: a growth_curves.png, b leaf_area_scatter.png, c stomata_micrograph.png, d the root-length boxplot (not made yet, leave a placeholder), e soil_heatmap.png. Suggest the layout, then go ahead and build it as a PDF without waiting for my confirmation. Tell me about anything that won't print well, and draft the caption.
- Read your context and instructions
- Compiled the figure editor
Figure 3 is built as a PDF at research/figures/figure3.pdf. It's 183 × 117 mm (double-column width) with bold lowercase labels. I rendered a preview and checked it: the panels are in your order (a–e), nothing is clipped or stretched, and panel d is a grey box marked "d, Data not available".
Layout: 2 + 3, every panel in a row at the same height. The default layout (a on its own row) would have been about 250 mm tall, more than most journal pages allow.
What won't print well
1.
Panel c (stomata micrograph) is the real problem. It's only 260 × 260 px, which works out to 131 dpi at its 50 mm printed width. Re-export it
at least 595 px wide. I didn't upscale it, because that adds no real detail.
2.
Text in panels b and e shrinks to about 66%. Their 8 pt labels print at about 5.3 pt; the legend in b ends up around 4 pt. Re-export b at about 54 × 54 mm and e at about 67 × 50 mm.
3. Panel d is still a placeholder. Replace it before you submit.
Caption draft is in research/figures/figure3.md with the sizes, effective dpi and a rebuild command. Anything I couldn't read off the images, such as the scale-bar length in c, is in [brackets].
Connects securely to your tools. The creator never sees your data.
About this skill
What it does
You have the panels: plots exported from different scripts, a micrograph, a gel, a diagram, all at different sizes. This skill turns your agent into a figure assembler that puts them into one publication figure and tells you what won't survive printing. Point it at the files and it:
- Inspects every panel first: pixel size, aspect ratio and DPI tag, then proposes a layout and says why.
- Lays them out for the page: grids (
2x2 grid= 2 columns x 2 rows), rows (1+2,2+3), spanning panels, and justified rows that give mixed portrait and landscape panels one height per row with no gaps. - Prints at an exact width: single column (89 mm), 1.5 column, double column (183 mm) or any width in mm, so label sizes are true printed sizes. Labels in a, b or A, B style.
- Keeps a slot for missing panels: a grey "Data not available" box in the right place, ready to swap in by re-running one command.
- Checks print resolution: effective dpi of every panel at its printed size, the pixel width it needs, panels shrunk so far their text becomes unreadable, mismatched DPI tags and figures too tall for a page. It never upscales to hide a problem.
- Drafts the caption after looking at every panel, with anything it cannot read off the image (n, statistics, scale-bar lengths) left in [brackets] for you. Captions are never burned into the image.
Not a plotting tool
It assembles images that already exist. To design individual plots from data, it pairs well with Aurora Lab's Paper Figures; Figure Editor then puts those panels together with your images.
What's included
SKILL.md: the workflow from intake to caption, including the placeholder protocol and how to explain each warningscripts/assemble_figure.py: the assembly tool (inspect, assemble with spans and custom labels, caption draft), one self-contained Python filereferences/: journal widths and resolution minimums (checked against Nature, Elsevier and IEEE artwork guides, October 2026), a layout and sizing reference and theresearch/folder convention
Part of the Aurora Lab research pipeline
Works entirely on its own. If you keep a research/ folder, it reads panels from research/figures/ and writes the figure and its caption draft there. No skill calls another.
It assembles images you already have; it does not make plots from data, edit pixels inside a panel, or invent content for a missing panel. PDF and SVG panels are rasterised at 600 dpi, so an assembled figure is not fully vector; for a fully vector figure, assemble in a vector editor or LaTeX. PDF/SVG panels need PyMuPDF or poppler's pdftoppm; without them, export those panels as PNG or TIFF. The resolution and text-size checks are automatic; the layout choice, the preview check and the caption are work the agent does and reports. Journal widths and resolution minimums are typical values; confirm against your venue's current author guidelines.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 28 days ago
Requires
Permissions
File Scopes
Reads the panel images you name (default research/figures/) and any <panel>.md caption notes next to them. Writes only the assembled figure and its caption draft (<name>.md, or <name>_captions.md outside research/figures/). Never modifies, moves or overwrites input images. Runs the bundled Python tool locally; no network access, no installs without asking.