Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Azzygoatcoder/agent-useful-skills --skill figure-drawinggit clone --depth 1 https://github.com/Azzygoatcoder/agent-useful-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/azzygoatcoder/agent-useful-skills/figure-drawing)<a href="https://agentmods.dev/skills/azzygoatcoder/agent-useful-skills/figure-drawing"><img src="https://agentmods.dev/badge/skills/azzygoatcoder/agent-useful-skills/figure-drawing/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/azzygoatcoder/agent-useful-skills/figure-drawing"><img src="https://agentmods.dev/badge/skills/azzygoatcoder/agent-useful-skills/figure-drawing.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00046 | $0.03186 |
| Opus 5 | $0.00023 | $0.01593 |
| Sonnet 5 | $0.00009 | $0.00637 |
| Haiku 4.5 | $0.00005 | $0.00319 |
Grade A, and why
figure-drawing scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Figure Drawing — 论文制图工作流
科研骨架的制图模块。目标是产出论文级（DAC/ICML 水准）矢量图。
第 0 步：场景判定（先定给谁看，再选路线）
科研制图分场景，先判定场景，再决定标准和工作流：
| 场景 | 触发 | 要求 | 标准 | 路线 |
|---|---|---|---|---|
| A. 内部理解 | 读别人论文精读、理解复杂架构、和 agent 对齐 | 中文可读、结构清晰、快速、可迭代 | 结构 9/10 就够，设计 7.5 可接受 | 直接画矢量（deepseek），不用 gpt-image-2 |
| B. 对外发表 | 自己的论文、投稿/审稿、graphical abstract、封面 | 英文、设计质量高、严格一致 | 设计系统等价、英文逐字、论文级 | gpt-image-2 出设计 → 复刻英文矢量 → 一致性检查 |
| C. 快速示意 | 组会 PPT、临时说明 | 最快、看懂就行 | 能表达即可 | mermaid 或直接画 |
| D. 数据图 | 精确数据/实验结果（折线/柱状/散点/误差棒） | 数据准确、可复现、期刊样式 | 数值与数据一致 + 视觉达标 | 代码驱动（pandas/matplotlib，bin/data_plot.py） |
例：LLSM 架构图 = 场景 A（读别人论文做理解图）→ 中文、直绘、不 gpt-image-2 是对的。
场景 B 补位（2026-08-11）：对外英文结构化图，尤其 sequence/state/ER/loop/radar/swimlane 等 draw.io 不擅长的类型 → 优先走独立 skill diagram-design（27 类型编辑器级，HTML+SVG 直绘，SVG 直接进 LaTeX
\includegraphics）。draw.io 场景仍用本 skill 时，套用"编辑级别设计系统"当自查表。技术/Agent 架构图补位（2026-08-12）：Agent/多智能体/系统架构图 → fireworks-tech-graph（语义形状：LLM=双边框圆角矩形、Agent=六边形、向量库=环柱体；Agent/记忆/RAG 领域模式内建；SVG 结构校验→PNG 视觉回读→定向修订的有界验证环）。与 diagram-design（editorial 排版）和 drawio（可编辑）定位不重叠——它专攻技术语义图。
工具不堆积原则（2026-08-12）：制图模块每个工具只占一个明确生态位（概念图=gpt-image-2 / editorial 图=diagram-design / 技术架构图=fireworks-tech-graph / 数据图=data_plot.py / 可编辑矢量=drawio）。新 skill 先判断：有新东西才吸收，重复轮子不安装——只把真正新增的能力/模式并进来，不平行堆工具。
核心工作流（5 步）
- gpt-image-2 生成概念图（按需，看选路规则）— 用 generate_image MCP（默认 provider）或详细 prompt。适合示意/机制/封面。
- prompt 用 PDCF 结构：类型 + 内容逻辑 + 风格hex + 负面限制（白底/扁平/≤3-4色组/无3D）
- 提取显式规格（透明化，防黑盒） — 用 Qwen3-VL 输出完整结构 JSON：所有节点/框（含逐字英文文本，绝不翻译/改写/简化）、箭头/连线（谁到谁+标签）、布局、视觉元素（图标/圆底数字等）。规格先展示给用户确认，缺什么提前指出——把"vision 读→写 XML"的黑盒变成可检查的显式规格
- 构造带样式 draw.io XML — 按下方 XML 约定写
.drawio（本 skill 自带规范）。严格英文逐字复刻（主流论文是英文，不翻译）。从规格构造，规格里有什么就画什么，不自行增删 - 验证（布局） — 用 drawio MCP 导出 PNG（
start_session→load_diagram→export_diagram；本机装了 draw.io CLI 也可draw.io --export --format png）→vision命令检查文字/布局/结构 - ✅ 一致性检查（必做，防丢失） — 把【原图 gpt-image-2 产物】和【矢量图导出 PNG】同时喂给 Qwen3-VL，显式要求：
"图A是原图，图B是矢量复刻。严格对比，逐条列出图B相比图A丢失/简化/改变的元素，重点关注文字省略、结构缺失、术语丢失。不要因为图B整洁就忽略差异。" 有差异 → 修
.drawio→ 重导 → 复检，直到差异最小化。
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 87 lines · 46 tokens per session scan A 92a8c4ca0639
figure-drawing is a skill published in the GitHub repository Azzygoatcoder/agent-useful-skills (6 stars, last pushed 11d ago), licensed MIT. It adds 46 tokens to every session and 3,186 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
drawio-academic-skills
Publication-figure overlay for draw.io. Use instead of drawio whenever the diagram is for a paper, thesis, dissertation, journal, conference, IEEE/ACM submission, manuscript, camera-ready, Word/LaTeX figure, or other publication. Applies venue, figure-type, color, caption/legend, formula, and paper-readability gates…
creating-figures
Creates publication-quality scientific figures for academic papers. Backend is chosen by host paper extension — .tex papers get TikZ figures, .typ papers get CeTZ figures; the two are never mixed. Use when creating block diagrams, system architectures, flowcharts, or technical illustrations for papers and theses.…
论文图表助手
A diagram-making assistant for flowcharts, system architecture diagrams, sequence diagrams, class diagrams, and mind maps. It can create standalone browser files that export diagrams as PNG images.
thesis-figure-skill
A skill for creating figures for academic papers, such as system diagrams, data-flow diagrams, and technical roadmaps. It can produce LaTeX/TikZ code or editable draw.io files, with support for compiling and checking the result.
product-design
A product-design method that turns a vague product idea into a small contract: one user problem, one output, and one decision. It applies to websites, apps, command-line tools, libraries, and other products.
creating-figures
Creates publication-quality scientific figures for academic papers using TikZ. Use when creating block diagrams, system architectures, flowcharts, or technical illustrations for papers and theses. Always render a temporary PNG/JPG preview and self-verify visual quality before finalizing figure files.