Run any Skill in Manus with one click

zoo-browser-verification

Stars5

Forks1

UpdatedJuly 8, 2026 at 16:49

Verify browser-visible behavior and capture focused UI evidence. Use after implementation for browser-impact work or when explicitly requested.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

andreyvit

andreyvit/zoo

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Software DevelopersComputer and Mathematical Occupations·SOC 15-1252

File Explorer

2 files

SKILL.md

readonly

name	zoo-browser-verification
description	Verify browser-visible behavior and capture focused UI evidence. Use after implementation for browser-impact work or when explicitly requested.

Follow .zoo/zoo.md, .zoo/browser.md if exists.

Prove browser-visible behavior works and collect actionable UI evidence.

Read task file, research file, acceptance criteria, current diff/change. Figure out flows to test.
Use repo browser/app harness mode when it exists and covers the needed flow.
Use harness in-page browser tool for navigation, interaction, inspection, screenshots, evidence. Codex: Browser Use. Claude Code: Chrome extension MCP (mcp__Claude_in_Chrome__*, see _ai/browser-testing.md).
Use OS-level automation only when in-page browser cannot exercise behavior: native alerts, print dialogs, permission prompts, frozen browser sessions. Codex Computer Use or Claude OS automation.
Save real screenshots under task evidence dir for each impacted flow or acceptance check, with realistic data, config, styling. Include empty/error states and expanded dropdown/popover/disclosure/select content.
Output acceptance checks with screenshot links proving pass.
Report pass/fail/blocked and why.

More from this repository

same repository

linus

andreyvit/zoo

Rant, swear, and insult like Linus Torvalds on the LKML — brutally blunt technical criticism. Use when user asks for Linus review/voice or for swearing.

2026-07-085

terse

andreyvit/zoo

Terse writing/comm style. Use for agent messages, skills, instructions, memories, docs, comments, commit messages and similar.

2026-07-085

zoo-add

andreyvit/zoo

Record user feedback for the active task without interrupting work. Use only when explicitly asked to add, remember, or handle later.

2026-07-085

zoo-cleanup-finished-specs

andreyvit/zoo

Archive completed Zoo spec files and resolved proposal files. Use when cleaning finished `.spec` work, moving completed specs into `archived/`, or moving implemented/rejected proposals into archive while honoring repo-local Zoo path overrides.

2026-07-085

zoo-code-review

andreyvit/zoo

Tiered code review gate: tests, scripted checks, narrow parallel checks, broad review. Invokable standalone on any diff or inside Zoo workflow.

2026-07-085

zoo-docs

andreyvit/zoo

Update durable documentation for completed Zoo work. Use when a Zoo docs writer or orchestrator should preserve practical learnings, behavior, or public documentation changes.

2026-07-085

name	zoo-browser-verification
description	Verify browser-visible behavior and capture focused UI evidence. Use after implementation for browser-impact work or when explicitly requested.

Follow .zoo/zoo.md, .zoo/browser.md if exists.

Prove browser-visible behavior works and collect actionable UI evidence.

Read task file, research file, acceptance criteria, current diff/change. Figure out flows to test.
Use repo browser/app harness mode when it exists and covers the needed flow.
Use harness in-page browser tool for navigation, interaction, inspection, screenshots, evidence. Codex: Browser Use. Claude Code: Chrome extension MCP (mcp__Claude_in_Chrome__*, see _ai/browser-testing.md).
Use OS-level automation only when in-page browser cannot exercise behavior: native alerts, print dialogs, permission prompts, frozen browser sessions. Codex Computer Use or Claude OS automation.
Save real screenshots under task evidence dir for each impacted flow or acceptance check, with realistic data, config, styling. Include empty/error states and expanded dropdown/popover/disclosure/select content.
Output acceptance checks with screenshot links proving pass.
Report pass/fail/blocked and why.