← 返回 Skills 市场
Scrapclaw
作者
Eric Pearson
· GitHub ↗
· v0.0.6
· MIT-0
277
总下载
0
收藏
0
当前安装
6
版本数
在 OpenClaw 中安装
/install scrapclaw
功能描述
Run Scrapclaw as a Dockerized browser-backed scraping service, then use this skill to fetch HTML from JavaScript-heavy or Cloudflare-protected pages through...
安全使用建议
This skill appears to do what it says: run/call a Dockerized Scrapclaw service. Before installing, (1) decide whether you'll use the published ghcr.io image or build from source — if building, review the Dockerfile and repository because docker-compose up --build can run arbitrary code. (2) Prefer running the container in an isolated VM or restricted host if you are unsure about target pages. (3) Confirm whether SCRAPCLAW_API_TOKEN must be set for your deployment — the SKILL.md treats it as optional but the registry metadata marks it required. (4) Do not expose the service to public networks and avoid scraping localhost/private RFC1918 targets unless explicitly allowed. (5) If you rely on the published image, verify the image source (ghcr.io/ericpearson/scrapclaw:v0.0.6) and its release notes before trusting it.
功能分析
Type: OpenClaw Skill
Name: scrapclaw
Version: 0.0.6
The scrapclaw skill is a legitimate tool for interfacing with a Dockerized browser-based scraping service. The SKILL.md file includes several proactive security instructions for the AI agent, such as warnings against accessing internal network ranges (SSRF protection), treating scraped HTML as untrusted input to prevent indirect prompt injection, and properly handling sensitive API tokens. No malicious behaviors, such as unauthorized data exfiltration or hidden execution of commands, were found.
能力评估
Purpose & Capability
Name/description describe a Dockerized browser-backed scraper and the skill only requires docker/curl/git and a base URL + optional API token — these align with running a container and calling its HTTP API. Note: git is only needed if building from source; the published image path does not require git.
Instruction Scope
SKILL.md contains concrete, scoped runtime instructions (health check, POST /v1 payload, cURL examples) and explicitly warns about reviewing Dockerfiles and avoiding internal-only targets. It does not instruct reading unrelated files or exfiltrating data.
Install Mechanism
Instruction-only skill (no install spec, no code files). The author recommends running an image from GitHub Container Registry or building from source; the doc correctly warns that building from unreviewed source can execute arbitrary host code.
Credentials
Requested environment variables (SCRAPCLAW_BASE_URL and SCRAPCLAW_API_TOKEN) are appropriate for calling the service. However, metadata lists SCRAPCLAW_API_TOKEN as required while SKILL.md treats it as optional (use only if set). Also primaryEnv is set to SCRAPCLAW_BASE_URL (a URL) which is unusual but not harmful. The declared required binaries include git even though using the published image does not need git.
Persistence & Privilege
Skill is not force-included (always:false), does not request persistent system-wide changes in the instructions, and does not modify other skills. Autonomous invocation is allowed (default) but not combined with other red flags.
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 scrapclaw - 安装完成后，直接呼叫该 Skill 的名称或使用
/scrapclaw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
v0.0.6
Add optional response truncation so callers can cap HTML or text payload size per request.
v0.0.5
Add text extraction mode, security hardening, improved install guidance, and ibarsi's contribution.
v0.0.4
Security hardening, improved install guidance, and ibarsi's contribution.
v0.0.3
Improve ClawHub description and install guidance
v0.0.2
Release 0.0.2
v0.0.1
Initial release
元数据
常见问题
Scrapclaw 是什么？
Run Scrapclaw as a Dockerized browser-backed scraping service, then use this skill to fetch HTML from JavaScript-heavy or Cloudflare-protected pages through... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 277 次。
如何安装 Scrapclaw？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install scrapclaw」即可一键安装，无需额外配置。
Scrapclaw 是免费的吗？
是的，Scrapclaw 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Scrapclaw 支持哪些平台？
Scrapclaw 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 Scrapclaw？
由 Eric Pearson（@ericpearson）开发并维护，当前版本 v0.0.6。
推荐 Skills