Hi Product Hunt, I'm Diogo. I built onsystem because coding agents write UI from their training data, not from your design system: palette colors instead of your tokens, native elements instead of your components, and APIs your design system doesn't have. onsystem reads your design system from source (components, props, variants, tokens) and gives it to the agent. In Claude Code, the plugin checks each file right after the agent edits it; Cursor, VS Code and Codex get it as an MCP server, and an ESLint plugin and a GitHub Action run the same check on every pull request. Try it on your app, nothing installed: npx -y onsystem check . I tested it in a 1,200-run benchmark across 6 models from 3 providers. Counted by onsystem 0.5.0's own check, design-system errors per run went from 0.85 without it to 0.00 with its skill, MCP server and hook, with no measurable change in tsc pass rate or cost. Scored by @shadcn/lint, a linter that is not part of onsystem, findings per run went from 4.97 to 2.64. Every run, the method and the caveats: https://onsystem.vercel.app/benc... It's MIT licensed and the roadmap is public. If you try it on a real project, tell me where it breaks.

onsystem

Keep coding agents on your design system

Keep coding agents on your design system