← 返回 Skills 市场
adcentury

Actionbook

作者 adcentury · GitHub ↗ · v0.1.1
cross-platform ⚠ suspicious
2070
总下载
0
收藏
16
当前安装
2
版本数
在 OpenClaw 中安装
/install actionbook
功能描述
Activate when the user needs to interact with any website — browser automation, web scraping, screenshots, form filling, UI testing, monitoring, or building AI agents. Provides pre-verified page actions with step-by-step instructions and tested selectors.
使用说明 (SKILL.md)

When to Use This Skill

Activate when the user's request involves interacting with a website:

Activate when the user:

  • Needs to do anything on a website ("Send a LinkedIn message", "Book an Airbnb", "Search Google for...")
  • Asks how to interact with a site ("How do I post a tweet?", "How to apply on LinkedIn?")
  • Wants to fill out forms, click buttons, navigate, search, filter, or browse on a specific site
  • Wants to take a screenshot of a web page or monitor changes
  • Builds browser-based AI agents, web scrapers, or E2E tests for external websites
  • Automates repetitive web tasks (data entry, form submission, content posting)
  • Wants to control their existing Chrome browser (Extension mode)

What Actionbook Provides

Actionbook is a library of pre-verified page interaction data. actionbook search finds actions matching a task description; actionbook get "\x3CID>" returns a structured document describing a page's purpose, functional capabilities, and DOM structure with inline CSS selectors — eliminating the need for runtime page structure discovery.

search and get

search — Find actions by task description

actionbook search "\x3Cquery>"                      # Search by task intent
actionbook search "\x3Cquery>" --domain site.com    # Filter by domain
actionbook search "\x3Cquery>" --url \x3Curl>          # Filter by URL
actionbook search "\x3Cquery>" -p 2 -s 20           # Pagination

Returns for each result:

  • ID — use with actionbook get "\x3CID>" to retrieve full details
  • Typepage (full page) or area (page section)
  • Description — page overview and function summary
  • URL — page where this action applies
  • Health Score — selector reliability percentage (0–100%)
  • Updated — last verified date

Constructing an effective search query

The query string is the primary signal for finding the right action. Pack it with the user's full intent — not just a site name or a vague keyword.

Include in the query:

  1. Target site — the website name or domain
  2. Task verb — what the user wants to do (search, book, post, filter, login, compose, etc.)
  3. Object / context — what they're acting on (listings, messages, flights, repositories, etc.)
  4. Specific details — any constraints, filters, or parameters the user mentioned (dates, location, category, language, etc.)

Rule of thumb: Rewrite the user's request as a single descriptive sentence and use that as the query.

User says Bad query Good query
"Book an Airbnb in Tokyo for next week" "airbnb" "airbnb search listings Tokyo dates check-in check-out guests"
"Search arXiv for recent NLP papers" "arxiv search" "arxiv advanced search papers NLP natural language processing recent"
"Send a LinkedIn connection request" "linkedin" "linkedin send connection request invite someone"
"Post a tweet with an image" "twitter post" "twitter compose new tweet post with image media attachment"
"Filter GitHub issues by label" "github issues" "github repository issues filter by label search issues"

When the user provides extra context (e.g., specific dates, a city name, a topic), fold it into the query even if it won't match a stored action literally — it helps the search engine rank relevant pages higher.

# User: "Help me apply for a software engineer job on LinkedIn"
actionbook search "linkedin job search apply software engineer application form"

# User: "I need to search for machine learning papers on arXiv"
actionbook search "arxiv advanced search papers machine learning subject category"

If --domain or --url is known, always add them — they narrow results and improve precision.

get — Retrieve full action details by ID

# Use the ID from search results directly
actionbook get "arxiv.org:/search/advanced:default"

Returns a structured document with:

  1. Page URL — exact URL and query/path parameters
  2. Page Overview — what the page does
  3. Page Function Summary — interactive capabilities (e.g., "Search Term Input", "Subject Classification Filtering")
  4. Page Structure Summary — DOM hierarchy with CSS selectors inline

Selectors appear embedded in the structure description, e.g.:

Search Term Form Section: Contains search term input field (input[type="text"]),
field selector dropdown (select[name="searchtype"]), and submit button (button.Search)

Extract CSS selectors from the structure summary for use with browser commands.

Browser Commands

Quick reference. Full details with all flags and options: command-reference.md.

Navigation

actionbook browser open \x3Curl>           # Open URL in new tab
actionbook browser goto \x3Curl>           # Navigate current page
actionbook browser back / forward       # History navigation
actionbook browser reload               # Reload page
actionbook browser pages                # List open tabs
actionbook browser switch \x3Cpage_id>     # Switch tab
actionbook browser close                # Close browser

Interactions

actionbook browser click "\x3Cselector>"          # Click element
actionbook browser fill "\x3Cselector>" "text"    # Clear and type
actionbook browser type "\x3Cselector>" "text"    # Append text
actionbook browser select "\x3Cselector>" "value" # Select dropdown option
actionbook browser hover "\x3Cselector>"          # Hover
actionbook browser press Enter                 # Press key

Observation

actionbook browser text                        # Full page text
actionbook browser text "\x3Cselector>"           # Element text
actionbook browser snapshot                    # Accessibility tree (live page structure)
actionbook browser screenshot                  # Save screenshot
actionbook browser screenshot --full-page      # Full page screenshot
actionbook browser wait "\x3Cselector>"           # Wait for element
actionbook browser wait-nav                    # Wait for navigation

actionbook browser close cleans up the browser session. Skip if the user requests the browser remain open.

Examples

User request: "Search arXiv for papers about Neural Networks, search in titles only"

# 1. Search — include the full intent: site + task + subject + filter preference
actionbook search "arxiv advanced search papers neural network title field" --domain arxiv.org

# 2. Get details — read Page Structure Summary for selectors
actionbook get "arxiv.org:/search/advanced:default"
# Response includes: input[type="text"], select[name="searchtype"], button.Search, etc.

# 3. Automate using selectors from the response
actionbook browser open "https://arxiv.org/search/advanced"
actionbook browser fill "input[type='text']" "Neural Network"
actionbook browser select "select[name='searchtype']" "title"
actionbook browser click "button.Search"
actionbook browser wait-nav
actionbook browser text
actionbook browser close

Fallback

Actionbook stores page data captured at indexing time. Websites evolve, so selectors may become outdated.

When a selector from actionbook get fails at runtime, actionbook browser snapshot provides the live accessibility tree with current selectors. Use selectors from the snapshot output to retry the interaction.

Selectors used in browser commands should come from actionbook get or actionbook browser snapshot output in the current session — not from prior knowledge or memory.

If actionbook search returns no results for a page, use snapshot as the primary source, or fall back to other available tools.

References

Reference Description
command-reference.md Complete command reference with all flags and options
authentication.md Login flows, OAuth, 2FA handling, session persistence
安全使用建议
This appears to be a coherent manual for a browser-automation CLI, not executable code — but there are important caveats: - Origin: the skill is instruction-only and references an external 'actionbook' CLI with no install provided. Verify the source and obtain the CLI from a trusted publisher before following commands. - Secrets: the docs show using environment variables for usernames, passwords, tokens, and cookies. Do not paste real credentials into example commands; prefer ephemeral credentials or test accounts. Confirm which env vars the agent or host will actually expose before running automation. - Persistence: profiles and cookies are stored on disk (e.g., ~/.config/actionbook/). If you use this, restrict directory permissions, delete profiles when finished, and avoid reusing profiles that contain unrelated account sessions. - Privileged operations: commands like browser eval and cookies set can access arbitrary page data and session tokens — run in an isolated environment or sandbox, especially for unfamiliar sites. - Metadata mismatch: the skill metadata declares no required binaries/credentials/config paths, but the instructions clearly use them. Treat that as a red flag: ask the publisher for clarification about expected local tooling, config locations, and what secrets (if any) you must supply. If you decide to use it: test against non-production/test accounts, run the CLI in a disposable container or VM, and wipe profiles/cookies after use.
功能分析
Type: OpenClaw Skill Name: actionbook Version: 0.1.1 This skill is classified as suspicious due to its extensive high-risk capabilities, which, while intended for legitimate browser automation, could be easily abused. Key indicators include the `actionbook browser eval "JavaScript code"` command (found in `references/command-reference.md`), allowing arbitrary code execution within the browser context, which is a significant remote code execution (RCE) risk. Furthermore, the skill provides comprehensive browser control, including full access to page content (`text`, `html`, `snapshot`), cookie management (`cookies set/get`), and the ability to connect to existing browser sessions (`--cdp`) or use custom browser executables (`--browser-path`). While the documentation in `references/authentication.md` provides security best practices, these powerful features create a substantial attack surface for data exfiltration, session hijacking, or other malicious activities if the AI agent is compromised or prompted maliciously.
能力评估
Purpose & Capability
The name/description (website interaction, scraping, form-filling, UI testing) align with the SKILL.md content: it is a manual for an 'actionbook' CLI that performs browser automation. However, the skill is instruction-only and provides many CLI commands that imply a separate 'actionbook' binary and local profile storage even though the registry metadata lists no required binaries or config paths — a usability/information-gap rather than a direct functionality mismatch.
Instruction Scope
The instructions explicitly teach how to log in to sites (including OAuth/SSO), fill credentials from environment variables (e.g., $APP_USERNAME, $GOOGLE_PASSWORD, $SESSION_TOKEN), set cookies, snapshot pages, eval JS, and persist profiles to disk. Those actions access and persist sensitive data and can reach arbitrary page content via eval; while expected for browser automation, the SKILL.md grants broad capability to read/manipulate credentials, cookies, and page data — and it references env vars and disk paths that are not declared in the skill metadata.
Install Mechanism
No install specification or code is included (instruction-only), so nothing is written to disk by the skill itself. This lowers supply-chain risk. Note: the instructions depend on an external 'actionbook' CLI whose provenance is unknown; the skill does not install it.
Credentials
The documentation routinely uses sensitive environment variables and tokens (APP_USERNAME, APP_PASSWORD, GOOGLE_PASSWORD, SESSION_TOKEN, ACTIONBOOK_API_KEY) and describes persistent profile directories and cookies, yet the skill metadata declares no required env vars or config paths. That mismatch obscures the fact that using this skill will involve handling secrets and persistent local session data — more sensitive than the metadata implies.
Persistence & Privilege
The docs instruct creating and reusing browser profiles (stored under a config path like ~/.config/actionbook/), persisting cookies and sessions, and recommend file-permission changes. The skill metadata does not declare these config paths or persistence behaviors. Persistent sessions increase the blast radius if misused; users should be aware that running the described commands will create long-lived artifacts on disk.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install actionbook
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /actionbook 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.1.1
Optimize the use cases and fallback handling
v0.1.0
actionbook 0.1.0 - Initial release of the actionbook skill for step-by-step interactive website automation and browser control. - Supports both dedicated browser (CDP) and Chrome Extension modes for flexibility with session and login handling. - Introduces Action Manuals: community-verified, pre-tested instructions with selectors for reliable automation. - Enforces a "manual first" workflow—search and retrieve an Action Manual for every new page type before acting. - Provides comprehensive CLI commands for searching manuals, extracting selectors, running browser actions, and managing extension setup.
元数据
Slug actionbook
版本 0.1.1
许可证
累计安装 17
当前安装数 16
历史版本数 2
常见问题

Actionbook 是什么?

Activate when the user needs to interact with any website — browser automation, web scraping, screenshots, form filling, UI testing, monitoring, or building AI agents. Provides pre-verified page actions with step-by-step instructions and tested selectors. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 2070 次。

如何安装 Actionbook?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install actionbook」即可一键安装,无需额外配置。

Actionbook 是免费的吗?

是的,Actionbook 完全免费(开源免费),可自由下载、安装和使用。

Actionbook 支持哪些平台?

Actionbook 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Actionbook?

由 adcentury(@adcentury)开发并维护,当前版本 v0.1.1。

💬 留言讨论