drawio-skill
Agents365-ai/drawio-skillSkill that turns natural language, Terraform, SQL, OpenAPI and other sources into editable draw.io diagrams, with incremental sync, drift detection and CI tests.
Skill that builds interactive isometric SVG, WebGL and 3D figures explaining a technical idea as a working physical object, then checks its own output.
Unusually thorough: a long, specific skill with a verification kit and many tests; expect long generations and a taste for detailed isometric style.
Anatomy makes the agent explain a concept by inventing a physical machine whose workings are the concept: a cache key becomes a pin-tumbler lock, a shared GPU context becomes a plotter with one print head moving across wells. The result is a framed card with title, hint and live readout, built from dozens of small shaded parts in four-tone shading, with calm eased motion that responds to the pointer and keyboard. Every number on screen comes from the same small model that moves the parts. The SKILL.md is very long (about 34 KB) and sits on top of a kit of 15 reference files, roughly 70 template files and about 100 script files, including Node scripts such as capture.mjs, drive.mjs, lines.mjs and turn-check.mjs that open the figure in Chrome or Chromium, zoom into joints and test interactions so the agent can catch its own mistakes. The repo has about 50 test files and CI, and the README says nothing needs installing beyond Node 20.10 or newer and a local Chrome, with no API keys. It is aimed at technical explainers and documentation visuals rather than general illustration.
A 34 KB SKILL.md, 15 reference files, about 70 templates, about 100 scripts, about 50 test files and CI, plus a self-check loop in headless Chrome.
Counted from the repository file tree. How we score depth.
A figure-drawing skill with a Node rendering kit and scripts that drive a local headless Chrome to capture, check and print the figures. The only code that leaves your machine is the optional publish step, which uses the GitHub CLI to push a branch and open a pull request after showing a plan and asking you to confirm; it also reads your Claude Code session log for time and token counts unless you pass --no-session. The README installs the skill by extracting the main-branch tarball into your skills folder.
The repository (173 files) on Oct 29, 2026. Not a full review: we read 7 of 106 code files; it relies on code outside the repository: Chrome or Chromium binary the capture/drive/print scripts launch, and the GitHub CLI and git used by publish --submit; The community gallery repository wheresryan22/anatomy-community that receives submissions; Most of the 106 code files were not opened (kit/*.mjs rendering libraries, examples/*, *.test.mjs, capture.mjs, drive.mjs, print.mjs, turn-check.mjs and others); a keyword search of them found only local-host CDP fetches and file writes.
NVIDIA SkillSpector static scan: 575 pattern hits (Memory Poisoning 483, analysis-evasion 72, MCP Rug Pull 8, Supply Chain 6, Rogue Agent 3, Agent Snooping 1). Of the 575 hits, 483 are Memory Poisoning notes and 72 are 'incomplete referenced artifact analysis' entries for files that SKILL.md links to and that exist in the repo. The obfuscation hits are large inline JSON DATA constants in generated example HTML pages. The '~/.codex/config.toml' hit is an install note in the README.
A review, not a certification. Skills change after we check them — read the files before installing.
MIT — commercial use allowed.
Not really. It is a creative skill with visual self-checks, but outputs need a person's judgment of the result.
Skill that turns natural language, Terraform, SQL, OpenAPI and other sources into editable draw.io diagrams, with incremental sync, drift detection and CI tests.
A skill that turns a description or a codebase into a validated, interactive standalone HTML diagram or one-page explainer.
A diagram skill that writes 45 diagram types as one self-contained HTML and SVG file in your brand, with onboarding, import and PNG export.
Agent skills from AntV's infographic engine: one writes a self-contained infographic from your text; four others help contributors extend its templates.
Anatomy is an agent skill that generates detailed, interactive isometric figures in SVG, WebGL and 3D. Each figure explains a technical idea as a physical machine that moves.
Install the skill, then ask for a figure, for example /anatomy explain how a token bucket rate limiter works, interactive. It needs Node 20.10 or newer and Chrome or Chromium for its checks.
No. The README says there are no API keys and nothing to npm install, since the kit uses only Node's built-in modules.
Those tools produce standard editable diagrams. Anatomy produces bespoke animated figures, which look more distinctive but cost far more tokens and are not editable in a diagramming tool.