Excalidraw

This prevents your main agent from wasting 4k-22k tokens reading verbose Excalidraw JSON files by delegating all diagram operations to subagents. Triggers automatically when you work with .excalidraw files or mention flowcharts and architecture diagrams. The problem is that Excalidraw files have terrible signal-to-noise ratios (a 79-element diagram is 22k tokens, mostly positioning and styling metadata, with actual content being maybe 10%). Without this, reading seven project diagrams burns 33% of your context budget. With delegation, subagents handle the JSON parsing and return compact text summaries, saving around 97% of tokens while keeping your main context clean for actual work.

skills.sh installs: 815
GitHub stars: 2.7K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add softaworks/agent-toolkit --skill excalidraw --agent claude-code

Installs into .claude/skills of the current project.

Files

Excalidraw Subagent Delegation

Overview

Core principle: Main agents NEVER read Excalidraw files directly. Always delegate to subagents to isolate context consumption.

Excalidraw files are JSON with high token cost but low information density. Single files range from 4k-22k tokens (largest can exceed read tool limits). Reading multiple diagrams quickly exhausts context budget (7 files = 67k tokens = 33% of budget).

The Problem

Excalidraw JSON structure:

Each shape has 20+ properties (x, y, width, height, strokeColor, seed, version, etc.)
Most properties are visual metadata (positioning, styling, roughness)
Actual content: text labels and element relationships (<10% of file)
Signal-to-noise ratio is extremely low

Example: 14-element diagram = 596 lines, 16K, ~4k tokens. 79-element diagram = 2,916 lines, 88K, ~22k tokens (exceeds read limit).