Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Slide Generation

v1.0.0

Generate Marp teaching slides from source content for teachers. Use when: (1) Generating Marp-compatible markdown slides from a .md file or folder for teachi...

0· 14·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for hh23333/slide-generation.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Slide Generation" (hh23333/slide-generation) from ClawHub.
Skill page: https://clawhub.ai/hh23333/slide-generation
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install slide-generation

ClawHub CLI

Package manager switcher

npx clawhub@latest install slide-generation
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
Name/description match the actual instructions: converting .md source files into Marp-compatible markdown slides using provided theme rules. The skill does not request unusual credentials or installs. Minor mismatch: SKILL.md / rules expect a source logo at assets/images/logo.png (and instruct copying it to output), but the file manifest does not include assets/images/logo.png; example pages contain remote image URLs which the skill's 'copy images' rule doesn't explicitly explain how to handle.
!
Instruction Scope
Runtime instructions tell the agent to read all source .md files, load references/*.md, create output folders, and copy any used images to the output images/ folder. That is coherent for a converter, but the instructions are vague about externally-hosted images (examples and theme files reference remote URLs) — copying such images would require network downloads. The SKILL.md also forbids adding external content ('Content must come ONLY from source files') while theme files include @import of external CSS/fonts; this creates ambiguity about allowed network fetches and what counts as 'source'.
Install Mechanism
Instruction-only skill with no install spec and no declared required binaries — minimal installation risk. Nothing in the package will be written to disk by an installer step prior to agent runtime (the agent's own actions will create output folders as per the instructions).
Credentials
No environment variables, credentials, or config paths are required. The skill does not request unrelated secrets or elevate privileges. This is proportionate to the stated purpose.
Persistence & Privilege
always is false and there is no install-time persistence or modification of other skills/configs. The skill relies on runtime file reads/writes within the project workspace, which is expected for a converter.
What to consider before installing
This skill appears to be a straightforward Marp slide generator, but before installing or running it: (1) verify that you will provide the source .md files and all referenced images locally (the skill expects to copy images to output). The manifest references a logo path (assets/images/logo.png) that is not present — either provide that asset or update the rules to avoid broken headers. (2) Decide whether the agent is allowed to fetch remote assets: theme SCSS imports external CDNs and some example slides reference remote images; the SKILL.md is ambiguous about downloading remote files. If you do not want the agent to access the network, ensure all images/fonts are included locally. (3) Test the skill in a sandbox or with non-sensitive content to confirm how it handles remote image links and missing assets. If you need, ask the author to clarify handling of externally-hosted images and to include the required logo asset or change the rules to not assume it exists.

Like a lobster shell, security has layers — review code before you run it.

latestvk978gj0zp1safjpwxr6m8wg5m185phrd
14downloads
0stars
1versions
Updated 5h ago
v1.0.0
MIT-0

Slide Generation Skill

Generate Marp slides from source content following the AM Blue Course theme and workflow rules.

Quick Start

  1. Identify sources: Read all source .md files (or all .md files in a folder)
  2. Read references: Load references/rules.md and references/marp-guide.md before generating
  3. Design structure: Plan slide sections from source content
  4. Generate: Output .md file with Marp format

Input Modes

Single File

User: 生成关于"记忆与检索"的幻灯片,基于第八章.md
→ Read 第八章.md, generate slides

Multiple Files / Folder

User: 把这个文件夹里的内容整合成一组教学幻灯片
→ Read all .md files in the folder, merge by topic/chapter, generate unified slides

When multiple sources are provided:

  • Merge content by topic or chapter sequence
  • Avoid duplicate content across sources
  • If sources conflict, prefer the more recent or comprehensive one

Output Target

When user specifies an output path (e.g. chapter8/memory-slide.md):

  1. Create the output folder (e.g. chapter8/)
  2. Copy all used images to chapter8/images/
  3. Write the slide markdown to chapter8/memory-slide.md

Image path mapping:

  • Source path: ../images/xxx.png → Output path: ./images/xxx.png
  • Copy source images to the output images/ folder alongside the slide file

Marp YAML Header (Required)

Always use this exact header:

---
marp: true
size: 16:9
theme: am_blue_course
paginate: true
headingDivider: [2,3]
footer: '*敬业乐群*'
---

Required Slide Structure

1. Cover Page (First Slide)

<!-- _class: cover_c -->
<!-- _paginate: "" -->
<!-- _footer: 敬业乐群 -->
<!-- _header: ![](images/logo.png) -->

###### [Subtitle]

# [Main Title]

@PresenterName
<Email>

Subtitle format: use ###### (level 6 heading) Main title: use # (level 1 heading)

2. Table of Contents (Second Slide)

## 本节内容

<!-- _class: toc_a -->
<!-- _header: "CONTENTS" -->
<!-- _footer: 敬业乐群 -->
<!-- _paginate: "" -->

- [Topic 1](#3)
- [Topic 2](#N)
- ...

3. Content Slides

Use ## (level 2 heading) for main section titles. The headingDivider: [2,3] setting automatically splits pages at ## and ### headings.

For content-heavy pages, apply text-size classes:

  • <!-- _class: smalltext --> — 0.9x font size
  • <!-- _class: tinytext --> — 0.8x font size

For narrow/wide images, avoid left-right layouts; use full-width instead.

Core Rules

Must Follow

  • ✅ Content must come ONLY from source files — do not add external knowledge
  • ✅ Keep content rich but within source bounds
  • ✅ Preserve original images from source; copy to output images/ folder
  • ✅ Use logo.png as header image (path: images/logo.png in output)
  • ✅ Presenter name and email as specified by user
  • ✅ Use <!-- *class: smalltext* --> or <!-- *class: tinytext* --> for dense content
  • ✅ Inline math: wrap in $...$ (e.g. $f(x) = y$)
  • ✅ Convert code to pseudocode if it aids comprehension
  • ✅ Section numbering: use 2.2 not 10.2.2; omit chapter prefix

Must NOT

  • ❌ Use classes not defined in theme files
  • ❌ Use transition slides (<!-- _class: trans -->)
  • ❌ Use --- for manual page breaks (headingDivider already handles this)
  • ❌ Title encoding like 10.2.2 — use 2.2 instead
  • ❌ Generate images — only use images from source content

Layout Decision Rules

Based on image/table shape:

Content ShapeRecommended Layout
Wide image (aspect < 1:1)Full-width, no columns
Tall image (aspect > 1:1)Left-right columns possible
Square imageFlexible, consider columns
TablesFull-width or right-aligned; avoid narrow columns
Code blocksFull-width, consider tinytext
Mixed text + imageUse cols-2 (五五分) or cols-2-64 (六四分)

Image scaling: use ![#c w:700](path) or ![#c h:300](path) to adjust size dynamically.

Column Layout Examples

Two columns (五五分):

<!-- _class: cols-2 -->
<div class=ldiv>

Left column content
</div>

<div class=rdiv>

Right column content
</div>

Two columns (六四分, image on left):

<!-- _class: cols-2-64 -->
<div class=limg>

![#c](image.png)
</div>

<div class=rdiv>

Right column text
</div>

fig-top layout (image on top):

<!-- _class: fig-top -->
<div class=fig-container>

![width:850px](image.png)
</div>

<div class=text-container>

Content text
</div>

Available Classes

Reference: assets/example_file/AwesomeMarp_blue.md and assets/example_file/slide.md

Page layout: cols-2, cols-2-64, cols-2-73, cols-2-46, cols-2-37, cols-3, rows-2, pin-3

Text size: tinytext, smalltext, largetext, hugetext

Special: toc_a, cover_c, fig-top, fglass, caption, fixedtitleA, fixedtitleB

Callouts: bq-purple, bq-blue, bq-green, bq-red, bq-black

Image Handling

  1. Scan source for ![]() image syntax
  2. Copy each image file to output images/ folder
  3. In slide markdown, reference images relative to output location: ![](images/filename.png)
  4. Map original paths (e.g. ../images/8-figures/8-1.png) to new paths (images/8-1.png)

For logos and theme images, use: images/logo.png (copied from assets/images/logo.png)

Multi-Source Synthesis

When synthesizing from multiple sources:

  1. Read all sources first — get full picture of content
  2. Identify overlap — note repeated topics across files
  3. Deduplicate — consolidate overlapping content
  4. Merge by structure:
    • If sources are chapters of the same book → follow chapter order
    • If sources are different perspectives → create unified section structure
    • If sources have conflicting info → prefer user's stated preference or most recent
  5. Generate one coherent slide deck — unified structure, consistent style

Quality Checklist

Before finishing, verify:

  • YAML header matches required format exactly
  • Cover page has all: main title, subtitle, presenter name, email, logo, footer, no page number
  • TOC page links to correct slide numbers
  • All images from source are included and copied to images/ folder
  • No Marp classes used beyond what's in theme files
  • No --- page breaks (headingDivider handles it)
  • Inline math uses $...$ syntax
  • No title encodings like 10.2.2 — use 2.2
  • Content stays within source material bounds
  • Logo path is images/logo.png

Comments

Loading comments...