← 返回 Skills 市场
sean810720

Search Strategy Skill

作者 https://github.com/account/ssh · GitHub ↗ · v1.1.0 · MIT-0
cross-platform ⚠ suspicious
303
总下载
0
收藏
1
当前安装
4
版本数
在 OpenClaw 中安装
/install search-strategy-skill
功能描述
智能搜尋策略執行器 - 自動選擇最佳搜尋工具,無需手動判斷。
安全使用建议
Before installing: 1) Verify the skill's provenance (the repo/homepage is generic), and consider pulling the code and reviewing it locally. 2) Expect to provide API keys (Firecrawl, Tavily, Brave, platform OAuth) even though metadata doesn't list them — treat those keys as sensitive. 3) Run the skill in a sandboxed environment or with restricted network access if you plan to pass untrusted URLs (to reduce SSRF/exfiltration risk). 4) If you don't need autonomous invocation, prefer to keep the skill user-invocable only. 5) Confirm whether you trust the external CLIs (firecrawl, tavily-search, agent-reach) before installing them; inspect or install them from official sources. 6) If using browser automation, explicitly check what BROWSER_SKILL_PATH points to and whether it can execute arbitrary Node code. If you want to proceed, add the required env var declarations to your policy and avoid passing internal-only URLs or secrets as query/url inputs.
功能分析
Type: OpenClaw Skill Name: search-strategy-skill Version: 1.1.0 The search-strategy skill is a legitimate utility designed to route search queries to various providers (Jina.ai, Firecrawl, Tavily, Brave, etc.) based on the task type. The shell script `scripts/search-strategy.sh` implements standard argument parsing and uses `curl` or external CLIs to fetch web content, with basic sanitization for URLs and queries. No evidence of data exfiltration, unauthorized credential access, or malicious prompt injection was found; the instructions in `SKILL.md` and `README.md` accurately describe the tool's intended logic and fallback mechanisms.
能力评估
Purpose & Capability
The name/description (search strategy) aligns with the code: it chooses among jina, multi (DuckDuckGo/Brave), firecrawl, tavily, and agent-reach. However the skill README and SKILL.md reference multiple API keys and a browser automation module (BROWSER_SKILL_PATH) that are not declared in the registry metadata (which lists no required env vars). That mismatch between declared requirements and actual usage is an inconsistency to be aware of.
Instruction Scope
Runtime instructions and the script routinely fetch arbitrary URLs (curl to user-supplied URLs, jina.ai proxy, search engine scraping, and optional Playwright automation). Fetching arbitrary URLs is expected for a web reader, but it also creates SSRF/side-effect risk if untrusted input is passed. The SKILL.md shows a node snippet referencing a BROWSER_SKILL_PATH and process.env.TARGET_URL which are not managed or validated by the script — this grants broad discretion over network access at runtime.
Install Mechanism
There is no install spec; this is instruction + script only, so nothing is downloaded at install time. That lowers install risk. The runtime does call other CLIs (firecrawl, tavily-search, agent-reach), but those are expected for the described functionality and must be installed separately by the user.
Credentials
The registry metadata declares no required env vars, but README and the script reference several credentials (FIRECRAWL_API_KEY, TAVILY_API_KEY, BRAVE_API_KEY, plus implied OAuth for agent-reach and BROWSER_SKILL_PATH). Required credentials are plausible for the advertised integrations, but the skill failing to declare them in metadata is a coherence issue and makes it easier to accidentally expose secrets. The script does not exfiltrate keys itself, but it will send user-supplied queries/URLs to external services.
Persistence & Privilege
The skill does not request 'always: true' and does not appear to modify other skills or system-wide settings. It runs as an invoked script and relies on other CLIs; autonomy is default but not escalated by extra privileges.
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 search-strategy-skill
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /search-strategy-skill 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
search-strategy-skill 1.1.0 - 移除了 SKILL.md 中的 metadata 欄位,提升文件簡潔性。 - 新增「最佳實踐」建議:fallback 到瀏覽器即代表網站需特殊處理,鼓勵用戶回報問題。 - 增加「未来規劃(Roadmap)」部分(內容未完待續)。 - 其餘核心架構、決策邏輯及用法說明保持原樣。
v1.0.2
**Major documentation and logic flow update for search-strategy-skill 1.0.2:** - Completely overhauled SKILL.md to provide clear technical design, decision-tree logic, error handling, and toolchain explanation. - Added detailed pseudocode and bash snippets for core algorithms, multi-level fallback, browser automation, and advanced search tool selection. - Improved documentation on parameter parsing, output formatting (including color), caching, and performance optimization strategies. - Clarified debug, logging, and monitoring methods for easier troubleshooting. - Included best practice recommendations and up-to-date usage tips. - No functional changes to the script logic were mentioned; documentation is now significantly more thorough and structured.
v1.0.1
search-strategy-skill v1.0.1 - Added SKILL_TEST.md and scripts/search-strategy.sh for improved testing and script execution. - Removed old deployment scripts (publish.sh, publish_check.sh). - Updated SKILL.md to the new structured format with explicit metadata and revised instructions. - Improved documentation clarity and structure.
v1.0.0
search-strategy-skill 1.0.0 - 初始發布:自動搜尋策略執行器,依任務自動選工具。 - 內建決策樹,支援單篇閱讀、完整爬蟲、多源比價、研究摘要等情境。 - 工具失效時可自動 fallback,無需手動切換。 - 提供統一介面,只需指定 `query` 或 `url`。 - 支援進階選項:強制選擇搜尋引擎、完整爬蟲等功能。 - 支援 markdown 與 JSON 兩種結果格式输出。
元数据
Slug search-strategy-skill
版本 1.1.0
许可证 MIT-0
累计安装 1
当前安装数 1
历史版本数 4
常见问题

Search Strategy Skill 是什么?

智能搜尋策略執行器 - 自動選擇最佳搜尋工具,無需手動判斷。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 303 次。

如何安装 Search Strategy Skill?

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

Search Strategy Skill 是免费的吗?

是的,Search Strategy Skill 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Search Strategy Skill 支持哪些平台?

Search Strategy Skill 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 Search Strategy Skill?

由 https://github.com/account/ssh(@sean810720)开发并维护,当前版本 v1.1.0。

💬 留言讨论