FigureCraft Skills 技能中心
标准化 Agent 插件生态架构。无缝打通 Claude Code、Cursor、Antigravity 与本地大模型。自动对齐 15 套顶刊数据契约（Data Contracts）与 5 维出版级 QC 质检规则。
交互式 Agent Prompt 生成工作台
选择技能并配置您的研究参数，1 秒生成适配 Claude / Cursor / Antigravity 的高精度系统提示词
选择 Agent 技能模块 (Built-in Skills)
5 个已安装技能 • 100% 离线可用Better Auth Security Best Practices
适用场景 (When to use)
<system> You are executing FigureCraft publication-grade scientific skill: `better-auth-security-best-practices` (v1.0.0). Your objective is to provide rigorous, publication-ready Python visualization and data engineering conforming to top-tier journal standards (Nature, Cell, Lancet, IEEE, The Economist). <skill_guidelines> Configure rate limiting, manage auth secrets, set up CSRF protection, define trusted origins, secure sessions and cookies, encrypt OAuth tokens, track IP addresses, and implement audit logging for Better Auth. Use when users need to secure their auth setup, prevent brute force attacks, or harden a Better Auth deployment. </skill_guidelines> <execution_constraints> 1. Always adhere to strict FigureCraft Data Contracts (JSON Schema column types, sample size N >= 3, null rate <= 5%). 2. Follow Ponytail minimalism: Pure Python standard libraries and native matplotlib/seaborn APIs; zero bloated dependencies. 3. Ensure 5-Pillar QC compliance: typography >= 8pt, dual export (SVG vector + 600 DPI raster TIFF/PNG), and colorblind accessibility. </execution_constraints> </system> <user> Execute `better-auth-security-best-practices` for the following scientific task configuration: Please output: 1. Complete, standalone reproducible Python script or analytical specification. 2. Data Contract schema requirements and validation notes. 3. Publication legend and styling rationale. </user>
官方内置技能全景索引 (Built-in Catalog)
覆盖从科研选题、数据清洗、绘图执行到质检验收的全流程
Better Auth Security Best Practices
Configure rate limiting, manage auth secrets, set up CSRF protection, define trusted origins, secure sessions and cookies, encrypt OAuth tokens, track IP addresses, and implement audit logging for Better Auth. Use when users need to secure their auth setup, prevent brute force attacks, or harden a Better Auth deployment.
Data Visualization
Create effective data visualizations with Python (matplotlib, seaborn, plotly). Use when building charts, choosing the right chart type for a dataset, creating publication-quality figures, or applying design principles like accessibility and color theory.
FigureCraft Curator
Expert chart selection and taxonomy curation assistant that maps experimental designs, variable types, and journal standards to FigureCraft publication templates.
- • When a researcher has tabular data but is unsure which chart best represents their statistical findings.
- • When planning figures for high-impact journals (**Nature, Cell, Lancet, IEEE, The Economist**) requiring specific visual conventions.
FigureCraft Data Wrangler
Data preprocessing and pandas transformation assistant that reshapes messy raw experimental datasets into strict FigureCraft Data Contract schemas.
- • When your raw laboratory data is formatted in wide matrices (e.g. one column per experimental group or replicate) rather than tidy long format.
- • When preparing datasets for FigureCraft engines that require strict column headers (`group`, `value`, `time`, `status`, `log2FoldChange`, `padj`, etc.).
FigureCraft Legend Writer
Academic figure legend and statistical methodology writer that generates publication-ready English captions tailored for Nature, Cell, Lancet, IEEE, and The Economist.
- • When drafting manuscript figure captions for journal submissions (**Nature, Cell, Lancet, IEEE, The Economist**).
- • When converting raw statistical analysis outputs into standard academic legend descriptions.
FigureCraft QC Audit
Automated pre-flight validation and human visual QC auditor enforcing the 5-pillar publication quality standard (typography >= 8pt, 600 DPI, accessibility, statistics).
- • Before submitting research figures to academic journals or bioRxiv/medRxiv preprints.
- • To audit generated vector files (`.svg`, `.pdf`) and raster assets (`.png`, `.tiff`) against editorial checklists.
FigureCraft Skill Creator
Scaffold generator and developer guide for authoring, packaging, and validating new FigureCraft agent skills and pluggable Python chart engines.
- • When creating a new FigureCraft agent skill for specialized bioinformatics, chemoinformatics, or economic visualization workflows.
- • When contributing a custom publication-grade chart engine to FigureCraft without modifying core application code.
Matplotlib
Low-level plotting library for full customization. Use when you need fine-grained control over every plot element, creating novel plot types, or integrating with specific scientific workflows. Export to PNG/PDF/SVG for publication. For quick statistical plots use seaborn; for interactive plots use plotly; for publication-ready multi-panel figures with journal styling, use scientific-visualization.
- • This skill should be used when:
- • Creating any type of plot or chart (line, scatter, bar, histogram, heatmap, contour, etc.)
Organization Best Practices
Configure multi-tenant organizations, manage members and invitations, define custom roles and permissions, set up teams, and implement RBAC using Better Auth's organization plugin. Use when users need org setup, team management, member roles, access control, or the Better Auth organization plugin.
Skill Manager 终端运维与 GitHub 技能安装
通过 scripts/skill_manager.py 轻松列出、校验、同步或从 GitHub 引入外部开源技能
1. 本地技能管理与规范校验
# 查看所有已安装技能
python scripts/skill_manager.py list
# 严格校验 SKILL.md 规范与脚本语法
python scripts/skill_manager.py validate skills/figurecraft-curator
# 搜索技能关键词
python scripts/skill_manager.py search "raincloud"
2. GitHub 外部技能导入与 Web 同步
# 从 GitHub 安装第三方技能 (内置 Zip Slip 安全防御)
python scripts/skill_manager.py install --github figure-craft/skill-single-cell
# 从本地 ZIP 或 URL 安全解压安装
python scripts/skill_manager.py install --url https://example.com/skill.zip
# 编译全量技能元数据至静态 Web 展厅
python scripts/skill_manager.py sync-web
..、绝对路径或跨盘符写盘的路径穿越攻击，确保执行环境绝对隔离安全。