← 返回 Skills 市场
donatasdecodo

Decodo Scraper

作者 DonatasDecodo · GitHub ↗ · v1.1.0
cross-platform ✓ 安全检测通过
1467
总下载
10
收藏
10
当前安装
3
版本数
在 OpenClaw 中安装
/install decodo-scraper
功能描述
Search Google, scrape web pages, Amazon product pages, YouTube subtitles, or Reddit (post/subreddit) using the Decodo Scraper OpenClaw Skill.
使用说明 (SKILL.md)

Decodo Scraper OpenClaw Skill

Use this skill to search Google, scrape any URL, or fetch YouTube subtitles via the Decodo Web Scraping API. Search outputs a JSON object of result sections; Scrape URL outputs plain markdown; Amazon and Amazon search output parsed product-page or search results (JSON). Amazon search uses --query. YouTube subtitles outputs transcript/subtitles. Reddit post and Reddit subreddit output post/listing content (JSON).

Authentication: Set DECODO_AUTH_TOKEN (Basic auth token from Decodo Dashboard → Scraping APIs) in your environment or in a .env file in the repo root.

Errors: On failure the script writes a JSON error to stderr and exits with code 1.


Tools

1. Search Google

Use this to find URLs, answers, or structured search results. The API returns a JSON object whose results key contains several sections (not all may be present for every query):

Section Description
organic Main search results (titles, links, snippets).
ai_overviews AI-generated overviews or summaries when Google shows them.
paid Paid/sponsored results (ads).
related_questions “People also ask”–style questions and answers.
related_searches Suggested related search queries.
discussions_and_forums Forum or discussion results (e.g. Reddit, Stack Exchange).

The script outputs only the inner results object (these sections); pagination info (page, last_visible_page, parse_status_code) is not included.

Command:

python3 tools/scrape.py --target google_search --query "your search query"

Examples:

python3 tools/scrape.py --target google_search --query "best laptops 2025"
python3 tools/scrape.py --target google_search --query "python requests tutorial"

Optional: --geo us or --locale en for location/language.


2. Scrape URL

Use this to get the content of a specific web page. By default the API returns content as Markdown (cleaner for LLMs and lower token usage).

Command:

python3 tools/scrape.py --target universal --url "https://example.com"

Examples:

python3 tools/scrape.py --target universal --url "https://example.com"
python3 tools/scrape.py --target universal --url "https://news.ycombinator.com/"

3. Amazon product page

Use this to get parsed data from an Amazon product (or other Amazon) page. Pass the product page URL as --url. The script sends parse: true and outputs the inner results object (e.g. ads, product details, etc.).

Command:

python3 tools/scrape.py --target amazon --url "https://www.amazon.com/dp/PRODUCT_ID"

Examples:

python3 tools/scrape.py --target amazon --url "https://www.amazon.com/dp/B09H74FXNW"

4. Amazon search

Use this to search Amazon and get parsed results (search results list, delivery_postcode, etc.). Pass the search query as --query.

Command:

python3 tools/scrape.py --target amazon_search --query "your search query"

Examples:

python3 tools/scrape.py --target amazon_search --query "laptop"

5. YouTube subtitles

Use this to get subtitles/transcript for a YouTube video. Pass the video ID (e.g. from youtube.com/watch?v=VIDEO_ID) as --query.

Command:

python3 tools/scrape.py --target youtube_subtitles --query "VIDEO_ID"

Examples:

python3 tools/scrape.py --target youtube_subtitles --query "dFu9aKJoqGg"

6. Reddit post

Use this to get the content of a Reddit post (thread). Pass the full post URL as --url.

Command:

python3 tools/scrape.py --target reddit_post --url "https://www.reddit.com/r/SUBREDDIT/comments/ID/..."

Examples:

python3 tools/scrape.py --target reddit_post --url "https://www.reddit.com/r/nba/comments/17jrqc5/serious_next_day_thread_postgame_discussion/"

7. Reddit subreddit

Use this to get the listing (posts) of a Reddit subreddit. Pass the subreddit URL as --url.

Command:

python3 tools/scrape.py --target reddit_subreddit --url "https://www.reddit.com/r/SUBREDDIT/"

Examples:

python3 tools/scrape.py --target reddit_subreddit --url "https://www.reddit.com/r/nba/"

Summary

Action Target Argument Example command
Search google_search --query python3 tools/scrape.py --target google_search --query "laptop"
Scrape page universal --url python3 tools/scrape.py --target universal --url "https://example.com"
Amazon product amazon --url python3 tools/scrape.py --target amazon --url "https://www.amazon.com/dp/B09H74FXNW"
Amazon search amazon_search --query python3 tools/scrape.py --target amazon_search --query "laptop"
YouTube subtitles youtube_subtitles --query python3 tools/scrape.py --target youtube_subtitles --query "dFu9aKJoqGg"
Reddit post reddit_post --url python3 tools/scrape.py --target reddit_post --url "https://www.reddit.com/r/nba/comments/17jrqc5/..."
Reddit subreddit reddit_subreddit --url python3 tools/scrape.py --target reddit_subreddit --url "https://www.reddit.com/r/nba/"

Output: Search → JSON (sections). Scrape URL → markdown. Amazon / Amazon search → JSON (results e.g. ads, product info, delivery_postcode). YouTube → transcript. Reddit → JSON (content).

安全使用建议
This skill implements a Decodo-hosted web-scraping client and requires a Decodo auth token (DECODO_AUTH_TOKEN). Before installing: 1) Verify you trust Decodo (https://decodo.com) because requests and scraped content are sent to scraper-api.decodo.com using your token. 2) Do not store the token in a committed .env file in a public repo — keep it in your environment or a secure secret store. 3) Reconcile the registry metadata mismatch (the registry claims no env vars while SKILL.md and the script require DECODO_AUTH_TOKEN). 4) Install dependencies locally (pip install -r requirements.txt) and inspect the code if you plan to run it in an environment with sensitive data. 5) If you need to prevent autonomous network calls, consider disabling model invocation or only invoking the skill manually. Overall the skill appears coherent for its described purpose, but only proceed if you trust the external Decodo service and follow best practices for handling the token.
功能分析
Type: OpenClaw Skill Name: decodo-scraper Version: 1.1.0 The OpenClaw skill is designed for web scraping via the Decodo API. The `SKILL.md` and `README.md` provide clear, non-malicious instructions for the AI agent and user, with no evidence of prompt injection attempts to subvert agent behavior. The `tools/scrape.py` script uses `argparse` for secure command-line argument parsing and the `requests` library for API communication to `https://scraper-api.decodo.com`. It correctly handles the `DECODO_AUTH_TOKEN` environment variable for authentication. There is no evidence of data exfiltration beyond the stated purpose, persistence mechanisms, obfuscation, or suspicious dependencies in `requirements.txt`. The code's functionality is directly aligned with its described purpose.
能力评估
Purpose & Capability
Name/description, README, SKILL.md, and the included Python tool all consistently implement a Decodo-backed web scraper (Google, Amazon, YouTube subtitles, Reddit, universal pages). The declared runtime requirement (DECODO_AUTH_TOKEN) is appropriate. Small inconsistency: the registry metadata at the top of the report lists no required env vars/primary credential, but SKILL.md and the script both require DECODO_AUTH_TOKEN.
Instruction Scope
SKILL.md and tools/scrape.py only instruct the agent to read a single DECODO_AUTH_TOKEN (or a .env file in the repo root), construct JSON payloads, and POST to scraper-api.decodo.com. The script does not read unrelated system files, other environment variables, or modify other skills or system settings.
Install Mechanism
There is no install spec (instruction-only), and included requirements.txt lists only requests and python-dotenv. No external arbitrary downloads, obscure URLs, or archive extraction are present. The user must install Python deps themselves (pip install -r requirements.txt).
Credentials
The only secret required is DECODO_AUTH_TOKEN (Basic token used in Authorization header) which is proportional to the task. Note the registry metadata did not list this required env var while SKILL.md and the code do — this mismatch should be reconciled before trusting the registry entry.
Persistence & Privilege
Skill is not always-enabled and is user-invocable. It does not request persistent elevated privileges, does not modify other skills or global agent configs, and does not request unrelated credentials.
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 decodo-scraper
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /decodo-scraper 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
v1.1.0
Expanded support for new content types: now includes Amazon, YouTube, and Reddit scraping. - Added commands to scrape Amazon product pages and Amazon search results (outputs structured JSON). - Added the ability to fetch YouTube video subtitles using the video ID. - Added support for scraping content and listings from Reddit posts and subreddits. - Google search output is now a structured JSON object with clearly described sections. - Documentation improved with detailed usage examples and updated summary table.
v0.1.0
- Added homepage and credentials metadata fields to SKILL.md. - Declared DECODO_AUTH_TOKEN as a required environment variable.
v1.0.0
Initial release of the Decodo Scraper OpenClaw Skill. - Search Google or scrape any web page using the Decodo Web Scraping API. - "Search" outputs a JSON array of results; "Scrape URL" outputs markdown content. - Easy command-line usage with clear examples for searching and scraping. - Requires `DECODO_AUTH_TOKEN` authentication. - Errors output as JSON to stderr with process exit code 1.
元数据
Slug decodo-scraper
版本 1.1.0
许可证
累计安装 10
当前安装数 10
历史版本数 3
常见问题

Decodo Scraper 是什么?

Search Google, scrape web pages, Amazon product pages, YouTube subtitles, or Reddit (post/subreddit) using the Decodo Scraper OpenClaw Skill. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 1467 次。

如何安装 Decodo Scraper?

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

Decodo Scraper 是免费的吗?

是的,Decodo Scraper 完全免费(开源免费),可自由下载、安装和使用。

Decodo Scraper 支持哪些平台?

Decodo Scraper 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 Decodo Scraper?

由 DonatasDecodo(@donatasdecodo)开发并维护,当前版本 v1.1.0。

💬 留言讨论