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 skills add Rheinmir/setup --skill dym-browser-use-replgit clone --depth 1 https://github.com/Rheinmir/setupWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/rheinmir/setup/dym-browser-use-repl)<a href="https://agentmods.dev/skills/rheinmir/setup/dym-browser-use-repl"><img src="https://agentmods.dev/badge/skills/rheinmir/setup/dym-browser-use-repl/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/rheinmir/setup/dym-browser-use-repl"><img src="https://agentmods.dev/badge/skills/rheinmir/setup/dym-browser-use-repl.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00105 | $0.01553 |
| Opus 5.5 | $0.00042 | $0.00621 |
| Sonnet 5.5 | $0.00021 | $0.00311 |
| Haiku 4.5 | $0.00011 | $0.00155 |
Grade A, and why
dym-browser-use-repl scanned grade A with 0 findings 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 yesterday.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
dym-browser-use-repl
When to use
- Việc cần nhiều bước có điều kiện — bước sau phụ thuộc thứ nhìn thấy ở bước trước. Script Playwright buộc phải viết mù toàn bộ kịch bản trước khi chạy; sai một chỗ là chạy lại từ đầu, mất cả login lẫn state.
- Đã viết
.mjsverify 2–3 lần mà vẫn trượt phần tử → đổi sang REPL, xem trang thật rồi mới quyết. - Cần chạm
file://hoặclocalhost(claude-in-chrome chặn 2 scheme này).
KHÔNG dùng khi: chỉ cần chụp một ảnh, hoặc chỉ đọc console một lần.
Đó là việc của /playwright-verify — một file .mjs, một lần chạy, xong vứt. Rẻ hơn.
Số đo (máy user, 09/09/2026, cùng một trang file://)
| cách | thời gian một lượt | state giữa các lượt |
|---|---|---|
Playwright script (chromium.launch() → goto → title → close) |
1.1–1.9s (đo 3 lần; lần đầu cold-cache 3.38s) | mất sạch |
browser-harness lượt đầu (khởi daemon + new_tab + screenshot) |
~2s | — |
| browser-harness lượt sau | 0.138s | tab + cookie + scroll giữ nguyên |
Steps
- Cài một lần (yêu cầu Python ≥ 3.11 —
python3mặc định của macOS là 3.9.6, sẽ trượt):python3.13 -m venv ~/.bh-venv && ~/.bh-venv/bin/pip install \ "git+https://github.com/browser-use/browser-harness" ~/.bh-venv/bin/browser-harness --version # 0.1.13 (đo 09/2026) - Chuẩn bị một Chrome có CDP — KHÔNG mượn Chrome của user (macOS sẽ bật hộp thoại xin
phép remote-debugging và chiếm tab đang làm việc). Dựng riêng một Chrome for Testing:
Chưa có binary →CH="$HOME/Library/Caches/ms-playwright/chromium-1243/chrome-mac-arm64/Google Chrome for Testing.app/Contents/MacOS/Google Chrome for Testing" "$CH" --remote-debugging-port=9333 --user-data-dir=/tmp/bh-prof --headless=new --no-first-run about:blank &npx playwright install chromium(dùng chung cache~/Library/Caches/ms-playwright). - Mỗi hành động = một heredoc. Helper đã import sẵn, daemon tự bám vào tab cũ:
export BH="HOME=/tmp/bhh BU_CDP_URL=http://localhost:9333 $HOME/.bh-venv/bin/browser-harness" env $BH <<'PY' new_tab("http://localhost:3000/checkout") # lượt ĐẦU dùng new_tab, không phải goto_url wait_for_load() print(page_info()) PY env $BH <<'PY' print(js("[...document.querySelectorAll('button')].map(b=>b.innerText)")) PY - Bấm bằng toạ độ đọc từ DOM, không từ pixel ảnh đã co:
r = js("document.querySelector('#submit').getBoundingClientRect()") click_at_xy(r["x"] + r["width"]/2, r["y"] + r["height"]/2) - Sau mỗi bước, moi lỗi ảnh chụp không thấy:
errs = [e for e in drain_events() if e["method"] == "Runtime.exceptionThrown" or (e["method"] == "Runtime.consoleAPICalled" and e["params"].get("type") == "error") or e["method"] == "Network.loadingFailed" or (e["method"] == "Network.responseReceived" and e["params"]["response"]["status"] >= 400)] - Dọn:
pkill -f "remote-debugging-port=9333".
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.
- yesterday First seen · 89 lines · 105 tokens per session scan A 0255f9e80860
dym-browser-use-repl is a skill published in the GitHub repository Rheinmir/setup (11 stars, last pushed 20d ago), licensed MIT. It adds 105 tokens to every session and 1,553 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-07.
dym-browser-use-repl compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| dym-browser-use-replRheinmir/setup | Skill | 11 | 20d ago | MIT | A | 105 | Claude Code |
| cua-driverdavidondrej/skills | Skill | 4.3k | 2d ago | MIT | A | 68 | Claude Code, Codex |
| create-verification-skillget-bb/bb | Skill | 5.1k | today | MIT | A | 57 | Claude Code, Codex |
| webapp-testinganthropics/skills | Skill | 189k | yesterday | none | A | 35 | Claude Code |
| browser-qahashgraph-online/awesome-codex-plugins | Skill | 1.5k | yesterday | Apache-2.0 | A | 58 | Claude Code, Codex |
| test-electron-appPostHog/posthog-foss | Skill | 728 | today | MIT | A | 112 | Claude Code |
| smoke-testSkyvern-AI/skyvern | Skill | 23k | today | AGPL-3.0 | A | 50 | Claude Code, Codex |