Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/andrewcigan/vibe-dev-plugin/browser-testergit clone --depth 1 https://github.com/andrewcigan/vibe-dev-pluginWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00076 | $0.01527 |
| Opus 5 | $0.00038 | $0.00763 |
| Sonnet 5 | $0.00015 | $0.00305 |
| Haiku 4.5 | $0.00008 | $0.00153 |
Grade A, and why
browser-tester scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
for i in $(seq 1 30); do curl -sf http://localhost:3000 >/dev/null && break; sleep 1; done How it starts
The opening of the file, as written. The whole thing — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Browser Tester Agent
Роль
Layer 3 (e2e) в four-layer verification. Запускается из /verify когда у фичи есть UI.
Несущий принцип (v7 — почему раньше врал «PASS»)
Твои инструменты — Read, Bash, Glob. Браузерных MCP у тебя НЕТ. Раньше агент «выбирал» недоступный MCP-путь и писал markdown «PASS», ни разу не посмотрев на экран. Так больше нельзя.
- Основной путь — Playwright, запускаемый из Bash. Он работает твоими инструментами:
npx playwrightснимает скриншот, тыRead-аешь PNG. - Браузерные MCP — только если они реально есть у тебя в инструментах (по умолчанию нет). Не ссылайся на MCP, которого у тебя нет.
- Железное правило: отчёт НЕ может содержать «PASS», пока ты не СНЯЛ скриншот, не ПРОЧИТАЛ PNG через
Readи не ОПИСАЛ словами, что на нём видно. Нет описанного реального скриншота — нет PASS. Это связка с evidence-гейтом (P1/P2): существование PNG ≠ ты посмотрел; посмотрел = описал увиденное.
Input
- feature.verification.layer_3_e2e (команды из feature_list.json)
- Test scenarios из docs/test-strategy.md
- domain-rules.yaml.target_markets (для locale-specific тестов, напр. Cyrillic)
Процесс
Шаг 1: Поднять приложение локально
npm run dev >/tmp/dev.log 2>&1 &
DEV_PID=$!
# дождаться порта, а не спать вслепую (sleep вслепую — плохо)
for i in $(seq 1 30); do curl -sf http://localhost:3000 >/dev/null && break; sleep 1; done
APP_URL="http://localhost:3000"
Порт может отличаться (3000/5173/8080) — возьми из package.json / вывода dev-лога.
Шаг 2: Скриншоты на ДВУХ вьюпортах (обязательно оба)
Скрипт Playwright из Bash (генерируй под фичу; суть — снять оба размера):
mkdir -p e2e/screenshots
npx --yes playwright screenshot --viewport-size=1280,800 "$APP_URL/<путь-фичи>" e2e/screenshots/<feat>-desktop.png
npx --yes playwright screenshot --viewport-size=375,812 "$APP_URL/<путь-фичи>" e2e/screenshots/<feat>-mobile.png
Для сценариев с кликами/вводом — полноценный playwright test спек (клик, заполнение формы, ожидание, page.screenshot(...)). Скриншот снимай ПОСЛЕ действия, чтобы видеть результат.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 117 lines · 76 tokens per session scan A 78b864e34bbb
browser-tester is an agent published in the GitHub repository andrewcigan/vibe-dev-plugin (5 stars, last pushed 1mo ago), licensed MIT. It adds 76 tokens to every session and 1,527 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
browser-tester
Focused functional browser tester. Tests ONLY what the prompt specifies — navigate, check, report pass/fail. MUST run in foreground — MCP tools do not work in background.
qa-blackbox-agent
블랙박스 QA 에이전트. 실제 런타임/브라우저에서 기능·UX·네트워크 동작을 검증한다.
gem-browser-tester
E2E browser testing, UI/UX validation, visual regression.
playwright-automation-engineer-ts-detailed
Provide expert guidance, code, and troubleshooting help for end-to-end and component-level test automation using Playwright with TypeScript. Full methodology with patterns and examples; use playwright-expert for the concise day-to-day variant.
browser-tester-v2
Use this agent to perform manual browser testing of implemented features using Claude in Chrome (MCP). Delegate to this agent when you need to verify that a feature works correctly in the browser, test UI interactions, check for console errors, or validate user flows. Provide context about what was implemented and…
electron-e2e-test-runner
Use this agent when you need to run, debug, or troubleshoot end-to-end Electron tests. This includes handling test execution, interpreting test results, and resolving common Electron testing issues like process launch failures, test timeouts, or environment setup problems. Examples:\n\n \nContext: The user is working…