excalidraw

Delegate Excalidraw diagram analysis to subagents for text summaries.

Updated Mar 28, 2026
One-click install
npx skills add https://github.com/sandalsoft/agent-skills --skill excalidraw-sandalsoft
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: excalidraw
Source: https://github.com/sandalsoft/agent-skills/tree/main/excalidraw
Command: npx skills add https://github.com/sandalsoft/agent-skills --skill excalidraw-sandalsoft

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Excalidraw files are large JSON documents with high token costs and noisy metadata, which can exhaust context budgets if read directly in the main agent.

Core Features & Use Cases

  • Delegation of all Excalidraw operations to specialized subagents to isolate context consumption and prevent budget overruns.
  • Text-focused summaries of diagrams (components, relationships, and architecture) without exposing raw JSON.
  • Safe, isolated analysis workflow for architecture visualization tasks such as explaining, updating, or summarizing diagrams.

Quick Start

Provide an Excalidraw file and ask for a concise architecture summary delegated to a subagent.

Frequently Asked Questions about excalidraw

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze Excalidraw diagrams without exceeding token context limits?

To analyze Excalidraw diagrams without exceeding context limits, delegate the Excalidraw analysis tasks to specialized subagents. This isolates the large JSON consumption, returning concise text summaries of components and relationships to the main context.

Why does reading Excalidraw JSON directly cause context budget overruns?

Reading Excalidraw JSON directly causes context budget overruns because the files are large documents with high token costs and noisy metadata. Using a subagent delegation protocol prevents loading this raw JSON into the main memory.

What is the best way to explain architecture visualization flows from Excalidraw files?

The best way to explain architecture visualization flows from Excalidraw files is delegating the visualization task to a subagent. This returns clear component and relationship insights as a text summary without exposing raw JSON data.

Can I update Excalidraw architecture diagrams without loading the full JSON into memory?

Yes, you can update Excalidraw architecture diagrams without loading the full JSON into memory by applying the delegation protocol. This workflow delegates updates to subagents, ensuring the main context only processes text summaries.

Does the subagent delegation protocol support summarizing complex architecture relationships?

Yes, the subagent delegation protocol supports summarizing complex architecture relationships by providing text-focused summaries. It extracts component insights and relationship details safely, preventing direct exposure to noisy Excalidraw metadata.

What are the limitations of using subagents for Excalidraw visualization tasks?

A key limitation is the strict delegation protocol: you cannot read Excalidraw files directly in the main context. The workflow returns text summaries only, meaning raw JSON access and direct in-context editing are restricted.