Token导航 LogoToken导航TokenDH.com
研究检索需要联网clawhub未标认证来源可访问clear审计通过

marp-data-summarizermarp 数据总结器

Agent Skill

用于辅助数据整理、表格处理、CSV/Excel 分析、指标计算和图表准备。它适合让 Agent 清洗字段、汇总数据、发现异常、生成统计口径或把分析结果转成可读说明。使用时需要确认数据来源、字段含义和时间范围,避免把样本数据当全量事实;涉及敏感数据、导出文件或批量写回时,应先确认权限和脱敏边界。

总安装

3,612

周安装

152

GitHub Stars

1

下载量

1,265
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:marp-data-summarizer(marp 数据总结器)
来源仓库:https://github.com/fuller-stack-dev/marp-data-summarizer
安装命令:
openclaw skills install marp-data-summarizer
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install marp-data-summarizer

简介

marp-data-summarizer 将各类资料自动转换为 Marp 格式幻灯片。

  • 支持文档、日志、截图与代码摘要的多源输入整合。
  • 输出美观且结构清晰的演示材料,适用于汇报与知识传递场景。
  • 需注意原始数据质量,避免错误摘要误导最终展示内容。
  • 适用宿主包括 OpenClaw,接入前应确认版本、权限和运行环境要求。

SKILL.md

name
marp-data-summarizer
description
Turn docs, notes, data, logs, screenshots, and code summaries into polished Marp slide decks.
homepage
https://marp.app
metadata
{ "openclaw": { "emoji": "📊", "requires": { "bins": ["marp"] }, "install": [{ "id": "node-marp", "kind": "node", "package": "@marp-team/marp-cli", "bins": ["marp"], "label": "Install Marp CLI (node)" }] } }

Marp Slide Deck Generator

Overview

Generate polished Marp slide decks from raw source material, then render them into the format the user wants. Focus on turning notes, data, and mixed inputs into a presentation with a clear takeaway, not a transcript dump.

If the user does not specify an output format, default to a rendered .html slideshow and also keep the .slides.md source deck.

Input Scope

This skill should work from almost any information that can be summarized into a presentation, including:

  • notes, docs, reports, and memos
  • tables, CSV, JSON, YAML, and metrics
  • logs, transcripts, timelines, and research notes
  • architecture docs, design docs, plans, and RFCs
  • code snippets, API descriptions, and technical change summaries
  • screenshots, images, and mixed structured + unstructured inputs

Workflow

  1. Identify the audience, purpose, and takeaway.
  2. Determine the requested output format from the user prompt, target filename, or explicit extension.
  3. If no format is specified, default to .html.
  4. Classify the input: numeric data, tabular data, text notes, logs, timeline, comparison, screenshots, code summary, or mixed.
  5. Extract the few facts that matter most.
  6. Choose a slide arc: title, context, key insights, supporting visuals, recommendation, close.
  7. Build the source deck as Marp markdown, typically named *.slides.md.
  8. Add visual structure with charts, diagrams, timelines, callouts, and section slides where useful.
  9. Render the deck into the requested format with the installed marp CLI.
  10. Verify the rendered deck in its target aspect ratio and confirm that no slide content is cropped or vertically cut off.

Output Rules

  • Accept any source material that can be turned into a coherent presentation.
  • Supported rendered outputs are .html (default), .pdf, .pptx, notes text, single-slide .png / .jpeg, and multi-slide image export.
  • Unless the user explicitly asks for markdown only, produce a rendered artifact.
  • If the user does not specify a format, render .html by default.
  • Keep the Marp source deck as *.slides.md when rendering to another format.
  • If the user asks for multiple formats, render all requested formats from the same *.slides.md source.
  • If the user asks for an unsupported format, ask one short question or offer the closest supported option.
  • Use direct marp CLI commands for rendering.
  • Use a strong title slide, then 4 to 8 content slides by default.
  • Keep one idea per slide.
  • Convert tables into short bullets or focused comparison slides when possible.
  • Use numbers, percentages, trends, and deltas instead of vague summaries.
  • Prefer short headers, short bullets, and purposeful whitespace.
  • Include speaker-friendly structure if the user wants notes.
  • If a slide is close to overflowing, split it into two slides instead of shrinking it until it becomes hard to read.
  • Avoid combining large tables, multi-card layouts, and large callouts on the same slide unless you have confirmed the rendered slide still fits.

Visual Patterns

  • Use mermaid for flows, timelines, funnels, and simple system diagrams.
  • Use comparison tables only when they improve readability.
  • Use callout boxes for key takeaways.
  • Use section divider slides for long decks.
  • Use a consistent palette and avoid clutter.
  • Favor simple, high-contrast layouts over dense decoration.

Deck Shape

  • Title slide
  • Context slide
  • Insight slides
  • Visual proof slides
  • Recommendation or summary slide
  • Optional appendix

Quality Check

  • Does the deck answer the user’s question fast?
  • Is each slide visually scannable?
  • Are the main numbers and conclusions obvious?
  • Does the deck feel designed, not auto-generated?
  • Does every rendered slide fit fully in-frame with no clipped bottom content or hidden overflow?

Resources

See references/slide-patterns.md for slide archetypes, visual rules, and mapping guidance.

See references/output-formats.md for supported output formats, default behavior, and render commands.

references/

Use references for reusable guidance on slide structure and visual decisions.

assets/

No bundled assets are required for this skill.


Keep the deck concise, visual, and specific.

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

77.32%
按下载量换算978

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills