Better context > more context. Been running experiments on LLM cost:

- Vanilla prompt $0.71
- With Storybook MCP: $0.39

That’s 45% fewer tokens for the same job.

We cut @chromatic.com total accessibility violations by over 66% in the last 6 weeks. Here's how:

Over the years, we accrued a11y debt because it was everyone and no one's responsibility. Sure, we did VPATs but they were out of date the moment they were finished.

A11y lags behind because testing doesn’t match how we build software. Why?

- Audits live outside the dev workflow
- Page-level tests create noise
- CI/CD gets blocked by a ton of old issues

Join me at SquiggleConf to build an a11y pipeline using Storybook!

To give folks a sense of frontend testing, the tools involved, and what each of them is meant for.

We pieced everything together in a Frontend Testing Trophy. Each testing type is described in detail.

h/t to all the other testing shapes: pyramid, trophy, etc

🏆 Thinking about what the testing trophy looks like for frontend devs.

Stronger focus on testing components, a11y, appearance.

Less focus on E2E and unit.