← Back to Skills Marketplace
davidme6

Smart Web Search

by davidme6 · GitHub ↗ · v3.1.0 · MIT-0
cross-platform ⚠ suspicious
250
Downloads
1
Stars
1
Active Installs
6
Versions
Install in OpenClaw
/install smart-web-search
Description
Smart Web Search v3.1 - Intelligent search with Baidu (China) and Google (International) as primary engines. Real-time news, ad filtering, content de-toxicat...
Usage Guidance
Summary of things to check before installing: - Metadata mismatch: ask the publisher to confirm required binaries (SKILL.md uses web_fetch; _meta.json lists 'curl'; top-level registry says none). Ensure your agent environment provides the expected fetch tool. - Network/privacy: the skill will send user queries to external search engines (Baidu, Google/Startpage, DuckDuckGo, etc.). Treat searches as network requests — do not include secrets or sensitive PII in queries you don't want sent to remote servers. - Vague processing steps: ad-filtering, de-tox, source verification and AI summaries are described but not implemented in code here; verify how your agent will perform those steps and test on non-sensitive queries first. - Placeholder scripts: some docs contain placeholder or example scripts (monitor.sh calling https://api.example.com, TODOs in feedback_collector). Confirm these are inert or review them if you plan to run any included scripts. - Provenance: source/homepage is unknown; repository and author links in _meta.json point to a GitHub user. If you need stronger assurance, request the real repository URL and inspect any code there. - Safe testing: run the skill in a sandboxed/limited environment first, monitor outgoing network calls, and verify behavior matches expectations before using for sensitive tasks. If the author can confirm/correct the curl requirement and clarify the monitoring/feedback scripts, many of the concerns are likely just documentation sloppiness rather than malicious intent.
Capability Analysis
Type: OpenClaw Skill Name: smart-web-search Version: 3.1.0 The bundle is a highly detailed web search utility designed to aggregate and filter results from multiple Chinese and international search engines (Baidu, Google via Startpage, 360, etc.). It utilizes the standard 'web_fetch' tool to retrieve content and provides the AI agent with extensive instructions for ad filtering, 'de-toxication' (filtering spam and misinformation), and source verification. No evidence of malicious intent, data exfiltration, or unauthorized system access was found; the complex markdown instructions in SKILL.md are strictly focused on refining search result quality and safety according to the stated purpose.
Capability Assessment
Purpose & Capability
The name/description match the instructions: SKILL.md instructs fetching search engine pages (Baidu, Startpage/Google, DuckDuckGo, etc.) and performing de-duplication, ad-filtering, and summarization. This functionality would not legitimately require credentials or broad system access. However, metadata is inconsistent: registry listing at the top shows no required binaries, while _meta.json lists 'curl' under requirements. Also some docs state 'no external API' while the skill explicitly fetches external search engine pages — this is functionally coherent but the wording across files is inconsistent.
Instruction Scope
The SKILL.md is instruction-only and instructs the agent to call web_fetch(...) against many public search endpoints. It does not tell the agent to read local files or request secrets. The high-level tasks (ad-filtering, de‑tox, source verification, AI summaries) are specified conceptually but the document does not supply concrete, auditable code for those steps — meaning the agent's own logic (or other skills) will implement them. That vagueness grants the agent discretion on how to filter, summarize, and what to include in outbound requests, which is a scope/clarity concern but not manifestly malicious.
Install Mechanism
There is no install spec and no code files executed at install time; the skill is instruction-only. This minimizes install-time risk. (Files are documentation only; there is no download/extract/execute install step.)
Credentials
Declared requirements differ between sources: top-level 'Requirements' shows none, but _meta.json lists 'curl' in requirements.bins. The skill itself uses a web_fetch tool in instructions (not curl). There are no required environment variables or credentials, which is appropriate for a public-web-search skill. Still, the inconsistency about required binaries should be clarified before use.
Persistence & Privilege
The skill is not always: true, does not request persistent system-level changes, and is user-invocable. It does include guidance for feedback collection and monitoring scripts (placeholders) but nothing that forces permanent presence or modifies other skills.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install smart-web-search
  3. After installation, invoke the skill by name or use /smart-web-search
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v3.1.0
v3.1: Baidu (China primary) + Google (International primary). All English documentation. Advanced search syntax. Real-time news filtering.
v3.0.2
v3.0.2: Fixed English description on ClawHub, all metadata in English
v3.0.1
v3.0.1: English description and metadata, improved international support
v3.0.0
v3.0: 新增百度/谷歌搜索、广告过滤、内容去毒、来源验证、警告系统,引擎总数达 12+
v2.0.0
v2.0: 实时搜索支持、多引擎聚合、智能去重、结果摘要、时间过滤
v1.0.0
- Initial release of Smart Web Search, version 1.0.0. - Automatically selects between domestic (CN) and international (DuckDuckGo) search engines based on language and query type. - Supports both Chinese and English input, with smart engine switching for topics and context. - Includes configurable engine priority, backup engines, and results filtering. - Built-in retry mechanism if the primary engine fails. - No API key required; all engines are free to use.
Metadata
Slug smart-web-search
Version 3.1.0
License MIT-0
All-time Installs 2
Active Installs 1
Total Versions 6
Frequently Asked Questions

What is Smart Web Search?

Smart Web Search v3.1 - Intelligent search with Baidu (China) and Google (International) as primary engines. Real-time news, ad filtering, content de-toxicat... It is an AI Agent Skill for Claude Code / OpenClaw, with 250 downloads so far.

How do I install Smart Web Search?

Run "/install smart-web-search" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Smart Web Search free?

Yes, Smart Web Search is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Smart Web Search support?

Smart Web Search is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Smart Web Search?

It is built and maintained by davidme6 (@davidme6); the current version is v3.1.0.

💬 Comments