← Back to Skills Marketplace
255
Downloads
0
Stars
0
Active Installs
4
Versions
Install in OpenClaw
/install vibe-reading
Description
Intelligent book reading and analysis skill. When the user provides an EPUB or TXT book file, split it into chapters, perform deep analysis and summarization...
Usage Guidance
Before installing, note three things: (1) copyright/legal risk — the instructions push the agent to produce 'high-fidelity' condensed versions that preserve anecdotes, tone, and dialogue. That behavior can reproduce large portions of a book and might violate copyright unless the user owns the rights or the work is public-domain. Prefer using this skill only with public-domain texts or ones you have rights to. (2) SKILL.md is truncated and there is a small metadata mismatch: the skill is listed as instruction-only in the registry but the SKILL.md requests pip installs (ebooklib, beautifulsoup4). If you allow the skill to run, be prepared for it to install Python packages and create/read/write files in the agent's working directories (input/, chapters/, summaries/). (3) If you need stricter behavior (e.g., summaries that avoid long verbatim excerpts), ask for explicit constraints up-front or decline to install. If you want higher confidence about exact behavior, request the missing/truncated portion of SKILL.md and a clear install/runtime plan from the publisher before proceeding.
Capability Analysis
Type: OpenClaw Skill
Name: vibe-reading
Version: 2.0.1
The 'vibe-reading' skill is a legitimate tool for analyzing and summarizing EPUB and TXT books. It provides structured instructions for an AI agent to extract text using standard libraries (ebooklib, beautifulsoup4), identify chapters, and generate detailed Markdown summaries and a static HTML reader. The instructions explicitly include safety measures, such as forbidding the embedding of API keys in generated HTML files, and show no signs of malicious intent, data exfiltration, or unauthorized execution.
Capability Assessment
Purpose & Capability
Name/description align with required binary (python3) and the declared Python libraries (ebooklib, beautifulsoup4). Asking the agent to extract text, split into chapters, and save files is coherent for a book-analysis skill. However, core instructions that demand a 'high-fidelity condensed version' equivalent to reading the original and preservation of anecdotes/dialogue can push the agent toward reproducing large verbatim sections of the source; that is disproportionate for a 'summarization/analysis' purpose and raises copyright/legal concerns.
Instruction Scope
SKILL.md explicitly instructs creating and running a Python script (ebooklib + BeautifulSoup) and reading/writing many local files (input/, chapters/, summaries/) — this is within expected scope. Concerns: (1) the instructions actively encourage the output to mimic the original text's tone and preserve detailed anecdotes/dialogue, which can result in reproducing copyrighted material rather than summarizing; (2) the SKILL.md is truncated ('Phase Five: Output G …[truncated]') so some runtime expectations are missing, creating ambiguity and risk of unexpected behavior; (3) the instructions require discretion-heavy operations ('AI-driven decision making' and 'take the time needed'), which are vague and give broad agent autonomy over what to extract and how much to reproduce.
Install Mechanism
The registry lists this as an instruction-only skill (no install spec), but SKILL.md includes an 'install' section that lists pip packages (ebooklib, beautifulsoup4). Installing these PyPI packages is reasonable for EPUB parsing; both are common packages. The mismatch between the registry 'no install spec' and SKILL.md's install directives is an inconsistency to flag (it affects user expectations about whether packages will be installed automatically). No high-risk external download URLs are present.
Credentials
The skill requests no environment variables, no external credentials, and no privileged config paths — proportional to its purpose. It requires filesystem access to read the user-supplied file and write intermediate and output files, which is expected for a document-processing skill. There is no evidence the skill attempts to access unrelated secrets or remote endpoints.
Persistence & Privilege
always:false and user-invocable:true — normal for a skill. There are no instructions to modify other skills or global agent settings. The skill asks the agent to create local files and run python scripts under the agent, which is standard and limited in privilege.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install vibe-reading - After installation, invoke the skill by name or use
/vibe-reading - Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.0.1
Fix security scan flags: move pip deps to frontmatter install declarations, remove runtime pip install instruction, remove explicit bash tool reference
v2.0.0
Major rewrite: Remove all Gemini SDK dependencies, exec() calls, and undeclared API keys. Pure instruction-based skill now compatible with any LLM model (GLM, Claude, Gemini, etc.). No external API keys needed. Resolves security scan flags.
v1.1.0
Added GitHub star CTA and ClawHub-compatible metadata
v1.0.0
Initial OpenClaw release - model-agnostic book reading and analysis skill
Metadata
Frequently Asked Questions
What is Openclaw?
Intelligent book reading and analysis skill. When the user provides an EPUB or TXT book file, split it into chapters, perform deep analysis and summarization... It is an AI Agent Skill for Claude Code / OpenClaw, with 255 downloads so far.
How do I install Openclaw?
Run "/install vibe-reading" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Openclaw free?
Yes, Openclaw is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Openclaw support?
Openclaw is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Openclaw?
It is built and maintained by BillWang (@drbillwang); the current version is v2.0.1.
More Skills