Agent skill

Agent Browser

by Peiiii in Peiiii/nextclaw

Use Agent Browser for real browser navigation, dynamic pages, and web interaction; also use it as a distinct browser-based path when websearch is unavailable or fails.

MITAuto-check passedProductivity & Automation

Install Agent Browser

skills CLI
$ npx skills add Peiiii/nextclaw --skill agent-browser -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install Peiiii/nextclaw agent-browser --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/Peiiii/nextclaw.git skills-src && mkdir -p .claude/skills && cp -r skills-src/packages/nextclaw-core/src/features/agent/shared/skills/agent-browser .claude/skills/agent-browser && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
agent-browser
GitHub stars
260
Token cost
~815 tokens
SKILL.md length
179 words
Files
1
Skills in repo
70
Repo updated
First seen
Licence
MIT

At a glance

Use Agent Browser for real browser navigation, dynamic pages, and web interaction; also use it as a distinct browser-based path when websearch is unavailable or fails.

  • Works in 3 steps: 明确告诉用户这会安装第三方全局 npm 包并下载浏览器运行文件。 → 用户当前请求若没有明确授权安装,先征得同意,不要静默安装。 → 获得授权后执行
  • Tasks that involve Browser automation
  • SKILL.md covers 能力边界, 首次使用与就绪检查, 先读取当前版本指南 and 标准执行循环, plus 4 more sections
  • Calls npm

What it does

Agent Browser is an agent skill from Peiiii/nextclaw. Use Agent Browser for real browser navigation, dynamic pages, and web interaction; also use it as a distinct browser-based path when websearch is unavailable or fails.

Its SKILL.md is about 820 tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Productivity & Automation, covering Browser automation and Web search. The repository describes itself as: A human-centered long-term AI partner—not a task-centered assistant. The licence is MIT.

When your agent uses it

  • Tasks that involve Browser automation
  • Tasks that involve Web search

Example prompts

  • “/agent-browser”

Requirements

  • Node.js

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

  1. 明确告诉用户这会安装第三方全局 npm 包并下载浏览器运行文件。
  2. 用户当前请求若没有明确授权安装,先征得同意,不要静默安装。
  3. 获得授权后执行

What it can do on your machine

Read from SKILL.md and the folder at commit 973722e. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npm

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Agent Browser loads about 815 tokens when it runs. Until then it costs about 46 tokens; SKILL.md has 179 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~46
When it runs · the whole SKILL.md, loaded when a task matches
~815

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from Peiiii/nextclaw at commit 973722e, republished under its MIT licence (© Peiiii). 179 words, ~815 tokens.

Download SKILL.mdSave it as .claude/skills/agent-browser/SKILL.md (or your agent's skills folder).
name
agent-browser
description
Use Agent Browser for real browser navigation, dynamic pages, and web interaction; also use it as a distinct browser-based path when web_search is unavailable or fails.
description_zh
使用 Agent Browser 完成真实浏览器导航、动态页面读取和网页交互;当 web_search 未配置或失败时,也可把它作为独立的浏览器路径评估使用。

Agent Browser

Agent Browser 是 NextClaw 调用的外部浏览器自动化 CLI。本 skill 只提供检查、安装引导、使用方式和安全边界,不包含 Agent Browser 的代码、npm 包或浏览器二进制。

能力边界

  • web_search 适合开放式信息发现,Agent Browser 适合访问已知 URL、读取动态页面和执行真实网页交互。
  • 两者是独立能力。不要把 Agent Browser 的结果伪装成 web_search 结果,也不要宣称它能稳定替代搜索服务。
  • 本 skill 中的 Agent Browser 专指外部 agent-browser CLI;web_fetch、Chrome/Edge DevTools MCP 和 Browser Connector 是其他独立能力,不能冒充或混称 Agent Browser。
  • 当 web_search 未配置、未启用或请求失败,而浏览器仍可能推进任务时,先用一句话告诉用户将改用浏览器自动化,再继续本流程。
  • 遇到验证码、登录墙、地区限制或搜索结果明显无关时,不得声称已经搜到或验证过信息。

首次使用与就绪检查

先检查外部 CLI 是否存在:

bash
command -v agent-browser
agent-browser --version

如果 CLI 不存在:

  1. 明确告诉用户这会安装第三方全局 npm 包并下载浏览器运行文件。
  2. 用户当前请求若没有明确授权安装,先征得同意,不要静默安装。
  3. 获得授权后执行:
bash
npm install -g agent-browser
agent-browser install

安装后或怀疑环境异常时先运行只读快速诊断:

bash
agent-browser doctor --offline --quick

只有诊断表明需要修复且用户同意时,才运行可能重装浏览器或清理旧状态的 agent-browser doctor --fix。

本地模式不要求付费云浏览器,也不要求 Browserbase、Browser Use 等云 provider key。选择外部云 provider 时,才按对应服务的账户和计费规则处理。

先读取当前版本指南

Agent Browser 的命令会随版本演进。执行非简单任务前,先读取 CLI 自带且与当前版本匹配的核心指南:

bash
agent-browser skills get core

只有需要完整参数或复杂工作流时再使用:

bash
agent-browser skills get core --full

以这份版本匹配指南为准,不要只凭记忆猜测命令或复制过时示例。

标准执行循环

为每个任务使用独立、简短、仅含小写字母、数字和连字符的 session 名称,例如 nextclaw-research-1。同一任务的所有浏览器命令都带同一个 --session:

bash
agent-browser --session nextclaw-research-1 open https://example.com
agent-browser --session nextclaw-research-1 snapshot -i -u
agent-browser --session nextclaw-research-1 get title
agent-browser --session nextclaw-research-1 get url

核心循环:

  1. open 打开明确 URL。
  2. snapshot -i -u 读取可交互元素和链接;只读正文时可使用完整 snapshot 或针对元素执行 get text。
  3. 使用最新 snapshot 生成的 @eN ref 交互。
  4. 页面导航、提交或动态重绘后先等待,再重新 snapshot;旧 ref 已失效。
  5. 用页面标题、最终 URL、可见文本或预期状态验证结果。
  6. 从实际打开和读取的页面记录来源 URL,不把搜索结果摘要当成已核验来源。

任务完成、失败或准备回复用户前,关闭本任务的 session:

bash
agent-browser --session nextclaw-research-1 close

不要默认使用 close --all,它可能关闭其他并行任务的浏览器 session。

用浏览器进行信息发现

浏览器信息发现是受限制的替代路径,不等于搜索 API:

  • 已知目标站点或域名时,优先打开站点首页、站内搜索页或用户提供的 URL。
  • 不知道来源时,可以尝试公开搜索引擎,但必须确认结果页确实对应查询,并打开具体来源页核验。
  • 某个搜索引擎出现验证码、空结果或明显无关结果时,可以换一种查询或另一个入口重试一次。
  • 两种策略都无法得到可核验来源时停止,关闭 session,并如实说明浏览器路径的限制;不要无限重试。
  • 对时效性事实,必须检查页面发布日期、更新时间和事件发生时间。

交互与授权

  • 公开页面的只读访问、snapshot、文本读取和截图,可以在用户请求范围内直接执行。
  • 登录、输入凭据、提交表单、购买、发帖、发消息、上传、下载或修改远端数据前,确认用户请求已经授权对应动作;没有授权就先询问。
  • 不把密码、token 或其他秘密直接写进命令参数或 shell 历史。需要认证时,先读取当前版本 core skill 中的 auth vault / profile 指南。
  • 页面动作结果不明确时先验证,不要重复点击可能产生双重提交的控件。

故障处理

  • CLI 或浏览器启动失败:运行 agent-browser doctor --offline --quick,再根据结果决定是否需要完整 doctor。
  • 页面等待超时:检查当前 URL、snapshot、console/errors;同一动作最多换一种等待或定位策略重试一次。
  • 出现验证码或反自动化阻断:不要尝试绕过,改用用户可操作的登录浏览器、其他公开来源,或报告阻塞。
  • 任意错误路径都应尝试关闭本任务命名 session;若进程已被外部强制终止而无法清理,要在后续恢复时先用 agent-browser session list 检查遗留 session。

完成检查

回复用户前确认:

  • 实际执行过需要的浏览器动作,而不是只描述计划;
  • 结论来自已打开并读取的页面;
  • 没有把浏览器访问称作 web_search;
  • 有副作用的动作都在授权范围内;
  • 本任务命名 session 已关闭。

© Peiiii, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in packages/nextclaw-core/src/features/agent/shared/skills/agent-browser of Peiiii/nextclaw.

Open the folder on GitHubat commit 973722e

Compare with similar skills

Agent Browser next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Agent Browser compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Agent Browser this skillPeiiii/nextclaw260—~815Automated safety check: PassMIT
Agent BrowserTorrent-Vibe/Torrent-Vibe159—~196Automated safety check: PassNone
Use Tinyfishtinyfish-io/tinyfish-cookbook2.2k—~2.1kAutomated safety check: PassMIT
Browse Nownowledge-co/community185—~619Automated safety check: PassNone
Web Research Workflowyonatangross/orchestkit289—~988Automated safety check: NotesMIT
Web Access via Browser CDPeze-is/web-access9.1k4 repos~2.2kAutomated safety check: PassMIT

Similar skills

  • Agent Browser

    Torrent-Vibe/Torrent-Vibe

    A skill your agent uses when a page must be opened, clicked, or extracted via the agent-browser CLI.

    159 GitHub stars~196 tokensUpdated 12 days ago
    Productivity & AutomationAuto-check passed
  • Use Tinyfish

    tinyfish-io/tinyfish-cookbook

    Use TinyFish for web search, fetching URLs, reading pages, current information, source-backed answers, research, docs, pricing/product pages, extraction, scraping, and browser automation.

    2.2k GitHub stars~2.1k tokensUpdated 6 days ago
    Productivity & AutomationAuto-check passed
  • Browse Now

    nowledge-co/community

    Control the user's actual browser through the browse-now CLI when a task needs authenticated pages, dynamic interaction, form filling, screenshots, or other browser automation that web search cannot…

    185 GitHub stars~619 tokensUpdated today
    Productivity & AutomationAuto-check passed
  • Web Research Workflow

    yonatangross/orchestkit

    Unified decision tree for web research and competitive monitoring.

    289 GitHub stars~988 tokensUpdated today
    Productivity & AutomationAuto-check: notes
  • Routes every web task, from searching to logged-in browsing, through a tiered choice of search, fetch, curl or a real Chrome or Edge session driven over CDP.

    9.1k GitHub starsUsed in 4 repos~2.2k tokens
    Productivity & AutomationAuto-check passed
  • Reddit JSON Fetcher

    ykdojo/claude-code-tips

    Fetches Reddit posts, threads and search results as JSON through a browser session, using a DuckDuckGo redirect to get past Reddit's automated-access block.

    10k GitHub stars~1.4k tokensUpdated 13 days ago
    Productivity & AutomationAuto-check passed

More from Peiiii/nextclaw

All 70 skills in this repo
  • Nextclaw Skin Studio

    Peiiii/nextclaw

    A skill your agent uses when a user wants to browse, apply, inspect, create, refine, switch, or remove a NextClaw skin; wants a personal skin with arbitrary CSS, JavaScript, images, SVG, DOM…

    260 GitHub stars~1.8k tokensUpdated yesterday
    Auto-check passed
  • A skill your agent uses when a development task needs visible phase tracing, task-level or phase-level Token measurement, model/effort comparison, or a deterministic usage report or local dashboard…

    260 GitHub stars~802 tokensUpdated yesterday
    Auto-check passed
  • 当 NextClaw 产品更新后需要生成、替换、挑毛病或检查官网、GitHub README、用户文档或社交传播中的真实截图、AI 宣传视觉、整页 HTML 宣传预览、社区二维码等对外视觉资产时使用;也用于“更新截图”“重新截一批图”“做宣传页”“生成 campaign 页面”“视觉审稿”“五星挑刺法”或发布前检查视觉资产。普通站点布局开发或只写文章不触发。

    260 GitHub stars~801 tokensUpdated yesterday
    Auto-check passed
  • UI UX Pro Max

    Peiiii/nextclaw

    A skill your agent uses when the user wants professional UI/UX design guidance, design-system generation, UX review, or stack-specific frontend guidance through a bundled local UI/UX Pro Max dataset…

    260 GitHub stars~2k tokensUpdated yesterday
    Auto-check passed
  • Impeccable

    Peiiii/nextclaw

    A skill your agent uses when the user wants distinctive, production-grade frontend design, anti-generic AI aesthetics, UX critique, technical UI audits, or final polish through bundled Impeccable…

    260 GitHub stars~2.5k tokensUpdated yesterday
    Auto-check passed
  • Curate NextClaw skill resources, including OpenClaw and community sources.

    260 GitHub stars~479 tokensUpdated yesterday
    Auto-check passed

Questions about Agent Browser

What does Agent Browser do?

Use Agent Browser for real browser navigation, dynamic pages, and web interaction; also use it as a distinct browser-based path when websearch is unavailable or fails. Agent Browser is an agent skill from Peiiii/nextclaw. Use Agent Browser for real browser navigation, dynamic pages, and web interaction; also use it as a distinct browser-based path when websearch is unavailable or fails.

When should I use Agent Browser?

Agent Browser fits situations like: tasks that involve Browser automation; tasks that involve Web search.

How do I install Agent Browser in Claude Code?

Run `npx skills add Peiiii/nextclaw --skill agent-browser -a claude-code`. Or copy the skill folder (packages/nextclaw-core/src/features/agent/shared/skills/agent-browser in Peiiii/nextclaw) into .claude/skills/agent-browser in your project. Claude Code loads it when a task matches its description.

How do I install Agent Browser in Codex?

Run `npx skills add Peiiii/nextclaw --skill agent-browser -a codex`. Or copy the skill folder (packages/nextclaw-core/src/features/agent/shared/skills/agent-browser in Peiiii/nextclaw) into .agents/skills/agent-browser in your project. Codex loads it when a task matches its description.

Can I use Agent Browser in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Peiiii/nextclaw --skill agent-browser -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/agent-browser, .gemini/skills/agent-browser, .github/skills/agent-browser and .opencode/skills/agent-browser in your project.

What does Agent Browser need to run?

Going by SKILL.md and its folder, Agent Browser needs the command-line tools its instructions call (npm). Our summary lists: Node.js.

Does Agent Browser access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Agent Browser safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Agent Browser use?

Agent Browser is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Agent Browser use?

About 815 tokens (SKILL.md is roughly 3.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Agent Browser?

Skills that share tags, products or a category with Agent Browser: Agent Browser (Torrent-Vibe/Torrent-Vibe, 159 stars), Use Tinyfish (tinyfish-io/tinyfish-cookbook, 2.2k stars), Browse Now (nowledge-co/community, 185 stars) and Web Research Workflow (yonatangross/orchestkit, 289 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Agent Browser?

Peiiii (a GitHub user) maintains it in Peiiii/nextclaw, which has 260 GitHub stars. The repository holds 70 skills in this directory. The repository was last updated on October 7, 2026.

Source: Peiiii/nextclaw on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.