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 samttoo22-MewCat/OpenSoul --skill agent-browsergit clone --depth 1 https://github.com/samttoo22-MewCat/OpenSoulWrote 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/samttoo22-mewcat/opensoul/agent-browser)<a href="https://agentmods.dev/skills/samttoo22-mewcat/opensoul/agent-browser"><img src="https://agentmods.dev/badge/skills/samttoo22-mewcat/opensoul/agent-browser/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/samttoo22-mewcat/opensoul/agent-browser"><img src="https://agentmods.dev/badge/skills/samttoo22-mewcat/opensoul/agent-browser.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 | $0.00075 | $0.00923 |
| Opus 5 | $0.00037 | $0.00462 |
| Sonnet 5 | $0.00015 | $0.00185 |
| Haiku 4.5 | $0.00007 | $0.00092 |
Grade A, and why
agent-browser 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 9d 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.
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.
What it actually says
agent-browser
Rust-based headless browser CLI optimized for AI agents. Uses Chrome DevTools Protocol (CDP) with Chrome for Testing.
安裝後設定
安裝完成後需下載 Chrome 執行環境:
agent-browser install
# Linux 需要系統套件:
agent-browser install --with-deps
使用時機(觸發情境)
- 需要截取網頁畫面
- 需要與 JavaScript 渲染的頁面互動
- 需要填寫表單、點擊按鈕
- 需要分析頁面存取性結構(accessibility tree)
- 需要生成網頁 PDF
- 需要在頁面中執行 JavaScript
核心指令
導航與截圖
# 前往網址
agent-browser navigate --url "https://example.com"
# 截圖(返回 base64 PNG)
agent-browser screenshot --url "https://example.com"
# 生成 PDF
agent-browser pdf --url "https://example.com" --output page.pdf
存取性快照(結構分析)
# 取得頁面 ARIA tree(含 element refs 供後續操作)
agent-browser snapshot --url "https://example.com"
元素互動
# 用自然語言或 ARIA 名稱找元素(返回 ref)
agent-browser find --url "https://example.com" --query "search bar"
# 點擊(使用 snapshot 或 find 取得的 ref)
agent-browser click --url "https://example.com" --ref "ref_42"
# 輸入文字
agent-browser type --url "https://example.com" --ref "ref_42" --text "hello world"
# 填寫表單(select/checkbox/input)
agent-browser fill --url "https://example.com" --ref "ref_10" --value "option_text"
JavaScript 執行
agent-browser eval --url "https://example.com" --expression "document.title"
進階功能
# 網路攔截(模擬 API 回應)
agent-browser mock --url "https://example.com" --pattern "/api/data" --response '{"mocked":true}'
# 裝置模擬
agent-browser navigate --url "https://example.com" --device "iPhone 14"
# 多標籤管理
agent-browser tabs --list
工作流程建議
- 先用
snapshot取得頁面結構與 element refs - 用
find定位目標元素(若 snapshot 結果難以分析) - 用
click/type/fill執行互動 - 用
screenshot確認結果
注意事項
- 首次使用前須執行
agent-browser install下載 Chrome - 截圖回傳 base64 PNG,可直接嵌入 markdown
snapshot回傳完整 ARIA tree,比截圖更適合結構分析
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.
- 9d ago First seen · 125 lines · 75 tokens per session scan A 26a53c97fcfe
agent-browser is a skill published in the GitHub repository samttoo22-MewCat/OpenSoul (19 stars, last pushed 5mo ago), licensed MIT. It adds 75 tokens to every session and 923 once invoked, about $0.0004 per session on Opus 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 2026-08-30.
Other skills, from other repositories
pm-preview
A tool that previews Markdown product documents in a web browser and refreshes the view when the documents change. Markdown is a plain-text format commonly used for project documentation.
python-cola-skill
A guide to organising Python back-end applications with the COLA architecture, a layered way to separate business rules, application actions, and technical connections. It applies to Flask and FastAPI projects and uses ideas from domain-driven design, which models software around the business domain.
Agent Browser
A fast Rust-based headless browser automation CLI with Node.js fallback that enables AI agents to navigate, click, type, and snapshot pages via structured commands.
scrapling-official
Scrape web pages using Scrapling with anti-bot bypass (like Cloudflare Turnstile), stealth headless browsing, spiders framework, adaptive scraping, and JavaScript rendering. Use when asked to scrape, crawl, or extract data from websites; webfetch fails; the site has anti-bot protections; write Python code to…
python-package-management
Guide for managing packages in the Agent Framework Python monorepo, including creating new connector packages, versioning, and the lazy-loading pattern. Use this when adding, modifying, or releasing packages.
temporal-python-testing
Test Temporal workflows with pytest, time-skipping, and mocking strategies. Covers unit testing, integration testing, replay testing, and local development setup. Use when implementing Temporal workflow tests or debugging test failures.