Utility mermaid diagrams
Skill product-on-purpose/pm-skills/skills/utility-mermaid-diagrams
Teaches PMs to create syntactically valid mermaid diagrams by selecting the right diagram type for their communication need, following syntax validity rules, and validating before shipping. Covers all 15 mermaid diagram types with PM-relevant examples and a dual-lens navigation system.From its SKILL.md
npx -y skills add product-on-purpose/pm-skills --skill utility-mermaid-diagramsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
What its file declares
Copied from the file, not written here
The file declares its own license as Apache-2.0. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
6.3 KB, ~1.3k tokens by cl100k_base, as published. Nobody here has run it
Mermaid Diagrams
Create effective, syntactically valid mermaid diagrams for product documents.
When to Use
- Creating mermaid diagrams for PRDs, specs, roadmaps, or stakeholder presentations
- Choosing which of 15 diagram types fits a specific communication need
- Debugging mermaid code that won't render or renders incorrectly
- Reviewing diagrams for clarity, accuracy, and accessibility
When NOT to Use
- Exporting diagrams to image files (PNG/SVG) - that's a rendering tool concern
- Using non-mermaid diagramming tools (Figma, Lucidchart, draw.io)
- Creating purely decorative visuals with no informational purpose
- Producing a full customer-journey artifact (stages, touchpoints, emotional curve, pain points) rather than a standalone diagram - use
discover-journey-map, which composes this skill for its timeline or flowchart section
The Cardinal Rule
Don't diagram what a list can say.
Diagrams earn their place when they reveal relationships, branching, or flow that prose flattens. Before creating any diagram, ask:
Does this show branching, relationships, or flow that a list or table would flatten?
- Yes → proceed with a diagram
- No → use a numbered list, bullet list, or table instead
A 5-step linear process is a list. A 5-step process with two decision points and a retry loop is a diagram.
Diagram Selection Guide
| I need to show... | Use | Also consider |
|---|---|---|
| A decision or approval process | Flowchart | State |
| Multi-service or multi-party interactions | Sequence | Flowchart |
| Feature lifecycle or status transitions | State | Flowchart |
| Work stages or pipeline status | Kanban | State |
| Release or sprint timeline with dependencies | Gantt | Timeline |
| Version history or chronological milestones | Timeline | Gantt |
| 2D prioritization (effort/impact, risk/value) | Quadrant | - |
| Allocation breakdown or composition | Pie | Treemap |
| Problem decomposition or brainstorming | Mindmap | - |
| Domain models or data relationships | ER | Class |
| API or object contracts | Class | ER |
| System topology or infrastructure | Architecture | Flowchart |
| Flow quantities or budget allocation | Sankey | Pie |
| Hierarchical proportional data | Treemap | Pie |
| Trends or time-series metrics | XY-Chart | - |
For worked examples organized by PM task, see references/pm-use-cases.md.
For full syntax and options per type, see references/diagram-catalog.md.
Syntax Validity Principles
Six rules that prevent most rendering failures:
- Quote labels - Any label containing spaces, parentheses, brackets, colons, commas, or reserved words must be quoted with double quotes
- Escape special characters - Characters with mermaid or markdown meaning (
>,<,-at line start,#) need escaping or quoting - Declare before referencing - Define a node before using it in an edge; referencing an undeclared node causes silent failures in some types
- Respect limits - Each diagram type has a maximum node/participant count beyond which readability collapses (see
references/diagram-catalog.mdfor per-type limits) - Comment your intent - Use
%%comments to document non-obvious choices (why this layout direction, why this grouping) - Test before shipping - Paste into a mermaid renderer (mermaid.live, VS Code preview, or your target environment) and verify it renders correctly
For the complete syntax reference, see references/syntax-guide.md.
Instructions
- Identify what you're communicating - What relationship, flow, hierarchy, or proportion needs to be visible? Who is the audience?
- Apply the cardinal rule - Confirm a diagram adds value over a list or table
- Select a diagram type - Use the selection guide above, browse
references/pm-use-cases.mdby task, or browsereferences/diagram-catalog.mdby type - Plan the diagram - Fill out the planning worksheet in
references/TEMPLATE.md: purpose, audience, node inventory, type rationale - Write the mermaid code - Follow
references/syntax-guide.mdrules; start with the minimal syntax example fromreferences/diagram-catalog.mdand expand - Validate - Run through the quality checklist below
- Embed - Place the validated mermaid code block in your document
Output Contract
- Planning artifact: A completed diagram planning worksheet (
references/TEMPLATE.md) - Final output: A syntactically valid mermaid code block embedded in the target document
- Quality gate: All items in the quality checklist pass
Quality Checklist
- Diagram renders without error in target environment
- Cardinal rule satisfied - a list or table would not communicate this more clearly
- No linear sequences without branching, relationships, or hierarchy
- All labels with spaces or special characters are properly quoted
- Special characters escaped where needed
- Node/participant count within type-specific limits
- Colors are accessible (WCAG AA 3:1 contrast minimum, black text on light backgrounds)
- Color is never the sole differentiator - shapes and labels also distinguish elements
- Diagram has a descriptive title or surrounding prose context
-
%%comments document any non-obvious layout or grouping choices
References
| File | Purpose |
|---|---|
references/TEMPLATE.md | Diagram planning worksheet - fill out before writing mermaid code |
references/EXAMPLE.md | Worked example: PM creating 4 diagrams for a product launch |
references/diagram-catalog.md | All 15 diagram types: syntax, PM examples, limits, pitfalls |
references/pm-use-cases.md | PM task → diagram type mapping with mini worked examples |
references/syntax-guide.md | Complete syntax validity rules, escaping, styling, and validation checklist |
What ships with it: 7 files
93.7 KB alongside SKILL.md
evals/
- trigger-fixtures.json3.5 KB
references/
- diagram-catalog.md40.8 KB
- EXAMPLE.md12.4 KB
- pm-use-cases.md19.5 KB
- syntax-guide.md13.4 KB
- TEMPLATE.md2.2 KB
- HISTORY.md1.8 KB
Gives 2 of the 12 instructions most diagrams charts skills give in ~1.3k tokens
Counted across 397 of the 418 authors here whose files we hold, read 2026-09-06
- Quote labels containing special charactershere, and in 18 of 397, across 17 files
- Choose the right diagram type for the datahere, and in 15 of 397, across 13 files
- Create a detailed Mermaid diagramin 12 of 397, across 11 files
- Validate Mermaid syntax before outputin 11 of 397, across 9 files
- Wrap diagrams in fenced Mermaid code blocksin 11 of 397
- Read the reference file for the chosen diagram typein 10 of 397, across 6 files
- Label every connection descriptivelyin 8 of 397, across 7 files
- Use stateDiagram-v2 for state diagramsin 7 of 397
- Declare the diagram type on the first linein 7 of 397
- Keep one concept per diagramin 7 of 397
- Cap diagrams at 15 nodes and split over 20in 7 of 397
- Describe the diagram before generating itin 7 of 397
Said here and by no other author read
- Confirm the diagram shows branching, relationships, or flow
- Fill out the planning worksheet before writing code
- Escape characters with mermaid or markdown meaning
- Declare nodes before referencing them
- Comment non-obvious choices using %% comments
- Test the diagram in a renderer before shipping
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.