Figure Review
Audit a scientific figure for publication-readiness: colormaps, uncertainty, axis labels, caption completeness, and claim support. Use this whenever the user shares or references a figure, plot, panel, or colorbar for a paper, report, or brief and wants it checked, reviewed, or made publication-ready — even if they just say "does this figure work?" or "review my plot." Emits a per-criterion repor…
Install
$ agentstack add skill-dgilford-ai-science-toolkit-figure-review ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-dgilford-ai-science-toolkit-figure-review) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Colorblind reference (loaded at runtime)
```! cat ~/.claude/skills/figure-review/COLORBLIND.md 2>/dev/null || echo "(colorblind guide not found)"
## House style (loaded at runtime)
```!
cat ~/.claude/skills/figure-review/CC-STYLE.md 2>/dev/null || echo "(no house style configured — criterion 6 skipped)"Inputs
Accept any subset of: figure image, plotting code, caption, surrounding text claim. Mark cant-assess for any criterion that requires input not provided.
Criteria
1. Colormap Flag colormaps listed as unsafe in the colorblind reference above (jet, rainbow, red–green). Verify sequential vs diverging choice matches data type. Assess luminance contrast. Warn if color is the sole encoding channel (no redundant shape/line style).
2. Uncertainty Flag if a quantitative claim is made and no uncertainty representation is shown (CI, spread, error bars, ensemble range, shading). If no quantitative claim is present, mark pass.
3. Axes Labels and units on all axes. Scale (linear/log) appropriate for the data range. Ticks legible.
4. Caption Self-contained. Defines all symbols, abbreviations, line styles. States n, time period, data source.
5. Claim support The figure shows what the surrounding text asserts. Flag over-reach. Requires the text claim as input; otherwise cant-assess. Example over-reach: figure shows 2°C warming at one station; text claims "warming is accelerating across the region."
6. House style (only when --style is passed) Check colors against the CC palette above, graphic fonts (Effra/Bebas/Work Sans), and visual direction (minimal clutter, white space).
Report
[Criterion]: pass / flag / cant-assess — [detail; if flagged, specific fix] Example: Colormap: flag — Uses jet. Replace with viridis or cmo.thermal.
Does not
- Make aesthetic or branding calls beyond the six criteria.
- Rewrite plotting code unless explicitly asked after the report.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: dgilford
- Source: dgilford/ai-science-toolkit
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.