← 返回 Skills 市场
229
总下载
0
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install media-crawl
功能描述
通过本机 media-agent-crawler HTTP 服务搜集 B站/抖音/YouTube/知乎内容（不依赖 MCP 客户端安装）。当用户要搜集这些平台内容、并已在本机启动应用（默认 http://127.0.0.1:39002）时使用。
安全使用建议
This skill is coherent with its description and simply calls a local media-agent-crawler service using bundled scripts. Before installing/using it: (1) ensure you actually run and trust the local service at the default address; (2) do not set BIL_CRAWL_URL to an untrusted external endpoint (that would cause the skill to send requests to that host); (3) review the included scripts (they are plain JS/sh and only call the local service) and run them in a safe environment if you are concerned about running shell/Node commands; (4) if you don’t have the local service, the skill will fail rather than installing external code.
功能分析
Type: OpenClaw Skill
Name: media-crawl
Version: 1.0.6
The media-crawl skill bundle is designed to interface with a local media-agent-crawler service (defaulting to http://127.0.0.1:39002) to fetch content from platforms like Bilibili, Douyin, YouTube, and Zhihu. It provides both Bash and Node.js scripts (e.g., `crawl.sh`, `mcp_tool.mjs`) to perform REST and MCP (Model Context Protocol) calls. The code and instructions in `SKILL.md` are consistent with the stated purpose of media scraping and do not contain any evidence of data exfiltration, unauthorized execution, or malicious prompt injection.
能力评估
Purpose & Capability
Name/description state it talks to a local media-agent-crawler service; included scripts and SKILL.md implement exactly that (REST POST to /start-crawl and JSON-RPC POST to /mcp). No unrelated binaries, credentials, or services are requested.
Instruction Scope
Instructions direct the agent to execute shell/Node commands and send HTTP requests to the configured base_url (default localhost). This matches the stated purpose, but the skill will execute arbitrary local shell/Node commands provided in the package (the scripts are visible) and will send data to whichever base_url is supplied (see environment override).
Install Mechanism
No install spec; this is instruction/script-only. All code is bundled in the skill (mjs/sh scripts). No downloads or external installers are used.
Credentials
The skill requests no credentials and no config paths. It does honor an environment variable BIL_CRAWL_URL (documented) to override the service URL — reasonable, but this means a user-supplied URL could redirect the skill's network traffic to an external host if misconfigured.
Persistence & Privilege
always:false and no special persistence/privilege escalation. The skill does not modify other skills or system-wide config; it runs only when invoked.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install media-crawl - 安装完成后，直接呼叫该 Skill 的名称或使用
/media-crawl触发 - 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
v1.0.6
- 增加脚本 scripts/stdio_utf8.mjs，用于支持 Node.js 下标准输入输出的 UTF-8 编码处理
- 内联命令部分：PowerShell 示例前增加设置 UTF-8 标准输入输出的指令，防止中文输出乱码
- 文档细节调整，确保 Windows PowerShell 用户获取更好中文兼容体验
- 其余功能和接口保持不变
v1.0.5
media-crawl 1.0.5 (no file changes detected)
- No code or documentation changes in this version.
- All functionality, commands, and usage remain the same as the previous version.
v1.0.4
media-crawl v1.0.4
- Added Node.js (`.mjs`) and PowerShell (`.ps1`) versions of major scripts: `crawl`, `crawl_mcp`, `list_archives_mcp`, and `mcp_tool`.
- Provided full cross-platform script support across Windows (PowerShell), macOS, and Linux (Node.js/curl).
- SKILL.md updated with new invocation priority: prefers inline shell commands (no dependencies), then Node.js scripts, then bash scripts.
- All script instructions updated to `skills/scripts/` directory for consistency.
- Enhanced usage demos and clarified environment-specific recommendations.
v1.0.3
- Renamed skill to "media-crawler-local" and clarified that it collects B站/抖音/YouTube/知乎 content without requiring the MCP client, communicating directly with a local HTTP service.
- Added structured tool list for all platforms, with required and optional parameters clearly documented.
- Expanded usage documentation for REST and MCP endpoints, providing detailed bash script usage examples for different collection and query scenarios.
- Documented new workflow for health checking, collection, summarizing results, and retrieving detailed task data.
- Described robust error handling strategies for common service and input failures.
元数据
常见问题
社交媒体研究助手Skill 是什么？
通过本机 media-agent-crawler HTTP 服务搜集 B站/抖音/YouTube/知乎内容（不依赖 MCP 客户端安装）。当用户要搜集这些平台内容、并已在本机启动应用（默认 http://127.0.0.1:39002）时使用。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 229 次。
如何安装 社交媒体研究助手Skill？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install media-crawl」即可一键安装，无需额外配置。
社交媒体研究助手Skill 是免费的吗？
是的，社交媒体研究助手Skill 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
社交媒体研究助手Skill 支持哪些平台？
社交媒体研究助手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 社交媒体研究助手Skill？
由 梅花三十三（@sansan-mei）开发并维护，当前版本 v1.0.6。
推荐 Skills