pm-diagrams

When to use

Whenever a visual is needed - architecture, domain structure, user flow, entity states, a business process, a journey map, or a strategic framework diagram. Call at any phase, with a specific type or with no argument for a recommendation.

Note: state machines (per entity) and sequence diagrams (per feature) are generated automatically within pm-entity-registry and pm-feature-design respectively. Use this skill for standalone diagram needs or to regenerate/update a diagram independently.

---

Two rendering tools

Mermaid.js (primary - Claude Code workflow) Rendered inline in markdown, embedded directly in live registers and Feature Cards. Claude Code reads and understands Mermaid syntax natively.

Excalidraw (secondary - human communication) Hand-drawn visual style for team and stakeholder communication. Rendered inline in conversation, exportable to excalidraw.com.

---

The 17-type catalogue

Each type has a dedicated composition reference (references/[slug].md) that the skill reads before drawing - the element vocabulary, composition rules, canonical structure, anti-patterns, and rendering notes for that specific notation.

Four types (dependency, gantt, kano, swot) render an analysis owned by another skill (pm-features-list, pm-product-roadmap, pm-market-analysis). If that analysis doesn't exist yet, pm-diagrams routes there instead of inventing one.

---

How to invoke

/pm-diagrams                          # no argument - context-aware recommendation
/pm-diagrams bpmn                     # specific type by slug - skips the menu
/pm-diagrams erd
/pm-diagrams --agent                  # autonomous draft from existing artifacts

---

How it works

1. With a slug argument: skips straight to gathering inputs for that type.
2. With no argument: reads workspace state (state.json, existing artifacts) and detects "visual gaps" - e.g. entities without a state machine, a feature list with dependencies but no graph - then recommends the top 3 highest-value diagrams for the current phase.
3. Loads the composition reference (references/[slug].md) for the selected type.
4. Reads the relevant source artifact first (entities.md, feature_list.md, JTBD analysis, market analysis...), then asks the user for anything missing.
5. Generates the diagram per the reference's rendering notes (Mermaid or Excalidraw).
6. For Mermaid: embeds in the relevant document (or outputs standalone).
7. For Excalidraw: renders inline + offers export/checkpoint.

---

Dependencies

No hard dependencies. Works from whatever artifacts are available; for the four "render for" types it routes to the owning skill if no analysis exists yet.

Best inputs per diagram type:

• State Machine / ERD / Domain Overview: entities.md (pm-entity-registry)
• Sequence Diagram: Feature Card Section 3 context (pm-feature-design)
• Dependency Graph / Kano: feature_list.md (pm-features-list)
• Gantt: pm-product-roadmap output
• JTBD Forces: JTBD Analysis output (jtbd-building)
• SWOT: pm-market-analysis output

Related skills: pm-entity-registry, pm-feature-design, jtbd-building, pm-features-list, pm-product-roadmap, pm-market-analysis