← Back to Skills Marketplace
redf426

graphviz

by redf426 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
137
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install graphviz-skill
Description
Generate architecture diagrams, flowcharts, dependency graphs, state machines, and other visualizations from natural language descriptions using Graphviz DOT...
README (SKILL.md)

Graphviz Diagram Generator

Generate diagrams from natural language descriptions and return a clickable link to GraphvizOnline with the rendered preview.

When to Use

  • "Draw the architecture of ..."
  • "Make a diagram showing ..."
  • "Visualize the data flow between ..."
  • "Create a flowchart for ..."
  • "Show dependencies between ..."
  • "State machine diagram for ..."
  • "ER diagram of ..."
  • Any request involving diagrams, graphs, schemas, architecture visualization

Workflow

  1. If the user's description is vague, ask one clarifying question before generating
  2. Generate valid DOT code following the conventions below
  3. Show the DOT code in a fenced ```dot block
  4. Construct the URL: https://dreampuf.github.io/GraphvizOnline/# + URL-encoded DOT code
  5. Output the link on its own line so it's clickable

DOT Code Conventions

Graph type

  • digraph for directed graphs (architecture, flows, dependencies, state machines)
  • graph for undirected graphs (relationships, peer networks)

Layout direction

  • rankdir=LR for wide/horizontal layouts (architecture, pipelines)
  • rankdir=TB for tall/vertical layouts (flowcharts, hierarchies)

Node shapes by component type

  • box or component -- services, applications, modules
  • cylinder -- databases, storage
  • ellipse -- users, actors, external systems
  • diamond -- decision points (flowcharts)
  • folder -- directories, namespaces, packages
  • note -- annotations, comments
  • record -- structured data, ER entities (use \x3Cfield> syntax for fields)
  • doublecircle -- start/end states (state machines)
  • circle -- intermediate states

Grouping

  • Use subgraph cluster_name { label="Group Name"; ... } to group related components
  • Give clusters a light fill: style=filled; color=lightgrey;

Styling

  • Use fontname="Helvetica" on the graph for clean rendering
  • Color-code edges or nodes to distinguish component types when helpful
  • Keep labels concise -- full names on nodes, short verbs on edges
  • Use style=dashed for optional or async connections

Edges

  • Label edges with the relationship or protocol: -> with [label="HTTP"], [label="gRPC"], [label="publishes"]
  • Use style=dashed for async/optional connections
  • Use dir=both for bidirectional relationships

URL Construction

Construct the GraphvizOnline URL by URL-encoding the entire DOT source and appending it as a hash fragment:

https://dreampuf.github.io/GraphvizOnline/#\x3CURL-encoded DOT code>

The DOT code must be encoded with standard percent-encoding (spaces -> %20, newlines -> %0A, etc.).

Output Format

Always output in this order:

  1. Brief description of what the diagram shows (1 sentence)
  2. The DOT source in a fenced ```dot block
  3. The clickable GraphvizOnline link on its own line

Example output structure:

Architecture showing the API gateway routing to backend services:

digraph { ... }

Open in GraphvizOnline

Usage Guidance
This skill appears to do exactly what it says: produce Graphviz DOT and give a GraphvizOnline link. Before using it, avoid embedding secrets or sensitive internal details (API keys, passwords, full internal hostnames) in diagrams because the DOT source will be included in the generated URL fragment and could be shared or copied. Note: the GraphvizOnline link places the DOT in the URL fragment (hash) which typically isn't sent to the server, but the URL itself can be copied or posted elsewhere—so treat diagram contents as potentially shareable. If you need to diagram sensitive infrastructure, consider rendering locally or using a private renderer instead of a public GraphvizOnline preview.
Capability Analysis
Type: OpenClaw Skill Name: graphviz-skill Version: 1.0.0 The skill bundle provides instructions for an AI agent to generate Graphviz DOT diagrams based on user descriptions and provide a link to a legitimate third-party renderer (dreampuf.github.io/GraphvizOnline). The files (SKILL.md, README.md, _meta.json) contain no executable code, obfuscation, or instructions to access sensitive system data or exfiltrate information.
Capability Assessment
Purpose & Capability
Name/description match the runtime instructions: the SKILL.md only asks the agent to produce Graphviz DOT, show the DOT source, and construct a GraphvizOnline URL. No unrelated credentials, binaries, or system access are requested.
Instruction Scope
Instructions are narrow and explicit: ask a clarifying question if the prompt is vague, generate DOT following the provided conventions, display the DOT block, and create an encoded URL for dreampuf.github.io/GraphvizOnline. The instructions do not request reading files, environment variables, or contacting endpoints other than the documented GraphvizOnline URL construction.
Install Mechanism
This is an instruction-only skill with no install spec and no code files. Nothing is written to disk and no external downloads are requested.
Credentials
No environment variables, credentials, or config paths are required. The declared requirements are minimal and proportional to the task.
Persistence & Privilege
The skill is not always-enabled and does not request elevated persistence or configuration changes. It is user-invocable and operates within normal autonomous-invocation defaults.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install graphviz-skill
  3. After installation, invoke the skill by name or use /graphviz-skill
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Major update: Skill now generates Graphviz DOT from natural language and returns a clickable diagram link. - Outputs a DOT code block and a direct preview link to GraphvizOnline. - Applies intelligent conventions for graph type, layout, node shapes, and grouping based on user prompts. - Asks for clarification if the input description is too vague. - No longer requires local WASM execution or direct SVG output.
Metadata
Slug graphviz-skill
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is graphviz?

Generate architecture diagrams, flowcharts, dependency graphs, state machines, and other visualizations from natural language descriptions using Graphviz DOT... It is an AI Agent Skill for Claude Code / OpenClaw, with 137 downloads so far.

How do I install graphviz?

Run "/install graphviz-skill" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is graphviz free?

Yes, graphviz is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does graphviz support?

graphviz is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created graphviz?

It is built and maintained by redf426 (@redf426); the current version is v1.0.0.

💬 Comments