- Home
- Skills
- Technical Documentation
- Excalidraw Diagrams - Local PNG
Works with the AI Agents you already use
Excalidraw Diagrams - Local PNG
Guide AI agents from editable Excalidraw JSON to local PNG diagrams without a cloud account.
Free
Excalidraw Diagrams - Local PNG
Example session with this skill installed
Explain the included two-service Excalidraw scene and give the local rendering command. Do not install dependencies or claim a PNG has been rendered if you have not run the CLI.
- Read your context and instructions
- Compiled the excalidraw diagrams
Included editable source
The package includes references/minimal-scene.json. It contains two labeled service boxes and one directed arrow:
- rect-a is the left service box at (80, 80), size 200 x 100.
- text-a contains the label "服务 A".
- rect-b is the right service box at (480, 80), size 200 x 100.
- text-b contains the label "服务 B".
- arrow-a runs from the right edge of the first box to the left edge of the second box, with its arrowhead at Service B.
The JSON includes five distinct element IDs, a white background and no embedded image assets. The example's text elements use fontFamily 5; confirm the appropriate font mapping in the installed CLI version before assuming that setting is portable across versions.
Render when dependencies are available
From the skill directory, inspect the installed command first:
excalidraw render --help
Then, if supported by that CLI version:
excalidraw render references/minimal-scene.json --out minimal.png
Verification status and next step
The editable example's JSON structure and element IDs were checked during packaging. No CLI rendering was run for this demonstration, so there is no verified PNG to show. After rendering locally, open minimal.png and verify that both Chinese labels are visible, text stays inside the boxes, and the arrow points from Service A to Service B. Deliver both the editable source and the inspected PNG.
Connects securely to your tools. The creator never sees your data.
About this skill
A free, original wrapper skill maintained by haxsd / agent-skills contributors for creating structured diagrams with the external @excalidraw/cli. It focuses on architecture diagrams, flowcharts and labeled boxes and arrows, rather than artistic illustrations.
What it teaches
- Describe nodes, labels and connections before building the scene.
- Create editable Excalidraw JSON using a bundled minimal example.
- Check element IDs, geometry, arrow connections and text settings.
- Render locally and visually inspect the PNG before claiming completion.
- Keep the editable source alongside the rendered output.
Local-first workflow
The default workflow uses the CLI's local render command. It does not require an Excalidraw+ account, workspace API key, cloud upload or paid AI model. Complex scenes can be authored by the current agent; delegation is optional and depends on the host's available tools and user authorization.
Requirements and language
Node.js 22+, an installed browser supported by the CLI and @excalidraw/cli are external prerequisites. Dependencies are not bundled and must be installed only with permission. Instructions are in Simplified Chinese; the listing is English. Consult the installed CLI's render --help for version-specific behavior and font mappings.
What's included
SKILL.md, a minimal two-service scene, source metadata, standalone attribution and a copyright/distribution notice. The original skill text and scene are maintained in haxsd/agent-skills; the external CLI is separately MIT licensed and is not included. This is not an official Excalidraw product, and no endorsement is claimed.
Evidence and limits
The package's ZIP structure, local Markdown links, scene JSON, unique element IDs and text settings have been checked. A CLI render was not completed during this submission preparation. The example below explains the bundled scene; it is not a screenshot or a claim that rendering was successfully tested. Inspect actual PNG output before relying on any generated diagram.
The free listing does not change ownership of the original skill text or automatically place it under the external CLI's MIT licence.
Actual CLI rendering and CI/CD execution have not been validated for this release; the completed checks cover package structure, Markdown links and example JSON only. Requires external Node.js 22+, @excalidraw/cli and a compatible installed browser; dependencies are not bundled. CLI flags and fontFamily mappings can vary by version. Chinese label rendering must be inspected in the resulting PNG. Simplified Chinese instructions focus on structured diagrams, not complex illustration. No Excalidraw+ cloud workflow or paid AI service is needed by the default local path. This independently maintained skill is not an official Excalidraw product.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
2 uses
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 1 month ago
- Free with an Agensi account
Requires
Permissions
File Scopes
Reads the bundled scene and user-selected diagram source files; writes editable JSON/Excalidraw files and PNG output to a user-approved directory. Terminal/process execution and a local browser process are needed by the external CLI. Network access may be needed only for separately authorized dependency installation; default local rendering needs no cloud service, API key, GitHub account or administrator access. Dependencies are not bundled. Actual rendering remains unverified for this release.