agentleFS
natsukiumnatsukium/claude-code-figures-plugin/skills/diagrams/SKILL.md

diagrams

Skill14 starsChanged 31 days agoApache-2.0View on GitHub

How to write diagram code blocks and display math that the figures plugin draws as pictures in the terminal (mermaid, dot/graphviz, d2, svg, LaTeX math). Load before putting a diagram or a formula in a reply, or when asked to draw, chart, or visualize a flow, architecture, state machine, dependency graph, or sequence.

9 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use observability.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • OpenAISkill

    observability

    Vercel Observability expert guidance — Drains (logs, traces, speed insights, web analytics), Web Analytics, Speed Insights, runtime logs, custom events, OpenTelemetry integration, and monitoring dashboards. Use when instrumenting, debugging, or optimizing application performance and user experience on Vercel.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If a block fails, the first line of the renderer's error is shown in red under the reply.

    line 13
  2. Very specific:

    Write one edge per line. A chained edge with a label, such as A --> B -->|ok| C, is misparsed.

    line 41
  3. Very specific:

    Keep labels short and quote any label containing punctuation: A["parse (stage 1)"].

    line 43

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Diagrams and math that render inline

A fenced block tagged mermaid, dot (or graphviz), d2, or svg in a reply is drawn as a picture under the reply, and the code stays visible above it. Where the terminal cannot show images, or the block's renderer is not installed, only the code is shown, so the source should still read sensibly as text.

If a block fails, the first line of the renderer's error is shown in red under the reply. Fix the source and send the corrected block, rather than explaining the error.

Choosing a language

  • mermaid for flowcharts, sequence diagrams, state machines, class and ER diagrams.
  • dot for graphs with many nodes or edges, such as dependency or call graphs; Graphviz lays them out more cleanly than mermaid does.
  • d2 for architecture diagrams with nested containers.
  • svg only for a figure none of the above can express.

Math

Display math, $$ … $$, \[ … \], or a math block, is typeset with MathJax and drawn under the reply. Inline $…$ stays plain text, so put a formula that matters on its own as display math.

  • amsmath environments such as aligned and cases, and mathtools, cancel, color, braket, and \boldsymbol, work.
  • Each formula becomes its own picture, so keep a derivation in one aligned block instead of many separate ones.
  • A latex block containing \documentclass is treated as a document and left as code.

Writing mermaid for mmdr

Mermaid blocks are rendered by mmdr, a reimplementation of mermaid.js that does not accept every construct mermaid.js does. mmdc, if installed, takes over only when mmdr reports an error, and mmdr misparses some blocks without one, so write for mmdr:

  • Write one edge per line. A chained edge with a label, such as A --> B -->|ok| C, is misparsed; write A --> B and B -->|ok| C instead.
  • Keep labels short and quote any label containing punctuation: A["parse (stage 1)"].

Size

A picture is at most 120 cells wide and 30 rows tall by default. One that would shrink to under 60% is grown up to the terminal's height, and if it is still small the user sees a note pointing to /figures. A long chain laid out top to bottom still ends up tiny; prefer flowchart LR for long chains, and split a diagram of more than about 20 nodes into several.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.