figure-style

Publication-grade figure correctness and legibility rules. Load before drawing any plot and call `apply_figure_style()` — sets a role-mapped font-size ladder, outward ticks, frameless legends, and 300-dpi output. The skill is a checklist, not a house look: data fidelity (claim-titles tested against every row, excluded data never enters summaries), label economy (floor and ceiling), colour threading, chart-choice-by-data-shape, layout, and a render-then-verify QA loop (bbox collision + per-panel perceptual check). Ships helpers: focal_palette, bar_with_points, strip_with_median, end_of_line_labels, panel_letter, set_frame, panel_crops. For multi-panel figures load `figure-composer`; for whole-paper figure arc load `paper-narrative`.

Quality

85%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

77%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A dense, actionable rulebook with an exemplary render-then-verify loop and concrete parameters throughout. Its main weakness is that everything lives inline in SKILL.md — the long checklist and QA code would benefit from one-level-deep reference files so the overview stays light.

Suggestions

Split §6 chart-family guidance into a reference file (e.g. references/chart-families.md) and keep a one-line-per-family decision table in SKILL.md, with clearly signaled links.

Move the §9.1/§9.2 verification code into a scripts/ or references/qa.md file and keep only the workflow ('run bbox check → fix → re-save until clean → per-panel perceptual crop check') inline.

Add a short quick-start block showing a complete apply_figure_style() call plus one helper usage so the entry-point workflow is executable without reading the full checklist.

Dimension	Reasoning	Score
Conciseness	Terse, rule-per-line prose that assumes Claude's competence (e.g. 'Never draw filled bars on a log-scaled value axis') with no explanation of known concepts, but §8 restates rules already given in §3.3, §4.5, and §1.1, which is minor trimmable repetition.	4 / 5
Actionability	Concrete parameters throughout ('ax.margins(0.04)', 'wspace≤0.06', 'apply_figure_style(sizes=(8,7,6))') and two copy-paste-ready code blocks (§9.1 bbox check, §9.2 panel crops), but no complete end-to-end usage example for apply_figure_style() or the named helpers — a minor gap.	4 / 5
Workflow Clarity	Explicit sequencing ('Load it and call apply_figure_style() before any plot'; §9 'After fig.savefig(...), before you finalize') with an explicit validate→fix→retry feedback loop ('Fix (move, shorten, stagger) and re-save until clean') and a per-panel perceptual checklist.	5 / 5
Progressive Disclosure	The body is well-sectioned (§0–§9) but is a single 330-line monolith with no bundle files and no file references; §6 chart-family guidance and the §9 verification code are candidates for separate reference files that could be loaded on demand.	3 / 5
	Total	16 / 20 Passed

Description

92%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A dense, highly specific description that clearly states what the skill does, when to load it, and how it differs from sibling skills. Trigger coverage is strong but could add a few natural synonyms (graph, matplotlib, publication-ready).

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'sets a role-mapped font-size ladder, outward ticks, frameless legends, and 300-dpi output' plus a full checklist enumeration (data fidelity, label economy, colour threading, chart-choice, layout, QA loop) and seven named helper functions — comprehensive rather than minor-gapped.	5 / 5
Completeness	Explicitly answers both what (the full capability list and named helpers) and when ('Load before drawing any plot'), with concrete routing triggers for sibling skills ('For multi-panel figures load `figure-composer`').	5 / 5
Trigger Term Quality	Natural terms users would say are present ('figure', 'plot', 'chart', 'publication-grade', 'multi-panel'), but common synonyms and tool words like 'graph', 'matplotlib', or 'publication-ready' are missing, so coverage is good rather than comprehensive.	4 / 5
Distinctiveness Conflict Risk	Clear niche — publication-grade scientific figure correctness — and it actively routes multi-panel and whole-paper work to `figure-composer` and `paper-narrative`, minimizing overlap risk with closely related skills.	5 / 5
	Total	19 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: UnicomAI/wanwu
Path: configs/microservice/bff-service/configs/agent-skills/claude-science/figure-style/SKILL.md
Commit: a16301c

Reviewed: 1 day ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.