mattiaippoliti/tia-skills/excalidraw-diagram

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Create or edit native Excalidraw diagrams for overall architecture maps, module diagrams, workflows, and relationships, with editable sources and verified SVG or PNG exports. Use when Excalidraw output or a reusable diagram for a document is requested.

Install

npx vskill@latest install mattiaippoliti/tia-skills/excalidraw-diagram

SKILL.md

Excalidraw diagrams

Make the relationship visible, and keep the result editable. Deliver a native .excalidraw scene; render exports from that scene so the source and the visible artifact agree.

Write plainly

Use unslop for diagram labels, captions, document prose, and changes to this skill. Call the Skill tool with unslop when available; otherwise read the installed unslop/SKILL.md and apply it. Keep technical names that identify real components. Explain unfamiliar terms in the surrounding prose.

Choose the visual argument

Identify the reader's question and the distinction the diagram must make: ownership, execution order, authorization, data transformation, or deployment. Use layout to express that distinction. A process can be a sequence; ownership can be a tree; independent work can fan out; a retry can loop back. Consistent shapes are useful when concepts have the same role. Vary them when the meaning changes.

For an overall architecture request, start with a synthesis map. Read synthesis.md. Show the major functional areas, their contents, shared infrastructure, external dependencies and a few labeled connections. A left-to-right execution sequence answers a different question. Use separate detail diagrams for a turn, ingestion or another process.

When a reference image is supplied, study its visual grammar before drawing. Reuse useful conventions such as cross-cutting bands, nested areas, a legend and restrained color. Populate them with the actual system's components. Do not import a reference's vendors or capabilities into the system being explained.

For an existing system, inspect its implementation or authoritative specification before naming boundaries and arrows. Record source paths or references alongside the working diagram. Distinguish current behavior from a proposed design. A code excerpt or payload earns space only when it resolves the reader's question.

Choose the final display size before laying out the scene. For a document, reserve its actual image width and height. If labels become too small, split the diagram or shorten them before scaling everything down. Aim for at least 9 pt labels at the final printed size.

Build the scene

Read authoring.md for the skeleton format, bindings, and palette. Use semantic IDs, native shapes and text, and routed arrows. The renderer supports native scenes and compact skeletons; generators are appropriate for repeated structures.

Group movable components deliberately. Bind a node's label to its container, and bind arrows to the shapes they refer to. Native scene bindings must be reciprocal. Preserve existing IDs, group membership, files, links, and unrelated elements when editing an existing scene.

Use containment for ownership or scope, arrows for a directed relationship, and lines for undirected relationships. Define dashed-line meaning locally when it differs from the solid path. Put annotations outside busy routes. Avoid implying a transaction or ordering guarantee the implementation does not provide.

Render and inspect

Read rendering.md on the first render or when dependencies are missing. The local renderer produces .excalidraw, .svg, .png, and a machine-check report. It blocks external network requests during rendering. Its pinned dependencies live in scripts/package-lock.json.

Run structural checks and inspect the actual PNG. Check every diagram at its intended display size: readable labels, adequate padding, correct arrow endpoints, unobstructed paths, and correct grouping. Inspect complex sections at full size as needed. A successful JSON parse or renderer exit is not visual approval.

After a correction, render and inspect the affected diagram again. Finish when the visible relationships match the source evidence and no layout defect remains. For a document replacement, render the document too: a good standalone diagram can still break pagination or become unreadable when embedded.

Deliver

Provide the editable scene and requested document or exports. For several figures, use descriptive filenames and a single archive of the sources. Keep diagnostic JSON and temporary renders out of the final delivery unless requested. State any font substitution or unavailable renderer capability that affects the result.

The design reference and the changes made here are recorded in provenance.md.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/mattiaippoliti/tia-skills/excalidraw-diagram/badge)](https://verified-skill.com/skills/mattiaippoliti/tia-skills/excalidraw-diagram)

Something wrong with this skill or its page? Tell us.