# Verifying With Browser

> Confirms a built UI or web page actually works before declaring the task done. Use after building or changing any frontend, page, or app with a UI — a server starting is not proof the page renders correctly.

- Skill: `aaravriyer193/verifying-with-browser` (Agent Skill)
- Install (CLI): `npx skillmds@latest add aaravriyer193/verifying-with-browser`
- Raw SKILL.md: https://api.skillmd.com/api/skills/aaravriyer193/verifying-with-browser/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Web & Frontend
- Author: aaravriyer193 (https://skillmd.com/u/aaravriyer193)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/aaravriyer193/verifying-with-browser

---


# Verifying with browser

A server starting or a build succeeding is not the same as a page working. Before saying a UI task is done:

1. `browser` action "navigate" to the local URL — it runs inside the sandbox, so `localhost:PORT` works directly, no rewriting needed.
2. `browser` action "screenshot" to actually see it, or "extract" to check the rendered text/structure.
3. Check for obvious breakage: blank page, error text, missing content, broken layout.
4. If there's an interactive element central to the task (a form, a button, a flow), exercise it rather than just eyeballing the landing state.

If something's broken, fix it and re-verify — don't present a link to something you haven't actually looked at. Same bar as running a test suite before saying tests pass.