安全扫描
OpenClaw
安全
high confidenceAn instruction-only skill whose requested actions (web search, web fetch, generate a .md note) align with its stated goal and it does not request additional credentials or install anything.
评估建议
This skill is internally consistent and low-risk: it simply directs the agent to search reputable web sources, extract structured information, and save a markdown notes file. Before installing/using it, confirm that (1) your agent environment provides the web-search and web-fetch tools it expects (mimo_web_search/web_fetch) and that those tools are trusted; (2) you are comfortable with the agent fetching external pages (including YouTube transcripts and third-party review sites); (3) you will no...详细分析 ▾
✓ 用途与能力
The skill's name/description (deep-reading a book and producing teaching-quality notes) matches the runtime instructions: perform web searches, fetch authoritative pages, extract and synthesize content, and save a markdown file. Requiring web search/fetch operations is appropriate and expected for this purpose; there are no unrelated binaries, env vars, or config paths requested.
ℹ 指令范围
SKILL.md is explicit about using mimo_web_search and web_fetch to gather 5-8 authoritative sources (publisher pages, Douban/Goodreads, author talks, case studies) and then producing a structured .md file. These instructions stay within the declared purpose and do not instruct reading unrelated local files or environment variables. Note: the skill expects the agent to have web browsing/fetch capabilities and will fetch external pages (including YouTube transcripts); that is expected but has privacy/copyright implications (see user guidance).
✓ 安装机制
No install spec and no code files — instruction-only — lowest-risk installation surface. Nothing is downloaded or written to disk by an installer; the only file write is the produced markdown in the workspace, which is appropriate for the skill's function.
✓ 凭证需求
The skill declares no required environment variables, no credentials, and no config paths. This is proportionate: the workflow uses public web sources and does not ask for unrelated secrets. If the agent had to access paywalled resources, the skill does not request credentials — which is consistent (it would require user-provided files or credentials, not implicit ones).
✓ 持久化与权限
always is false and autonomous invocation is allowed by default — appropriate. The skill writes its own output file into the workspace (expected). It does not request persistent system-wide privileges or modify other skills' configurations.
安全有层次,运行前请审查代码。
运行时依赖
无特殊依赖
版本
latestv1.0.02026/4/16
初始版本:深度读书笔记生成技能,支持五层分析框架,可生成教学级别的书籍知识体系笔记
● 无害
安装命令
点击复制官方npx clawhub@latest install book-deep-reader
镜像加速npx clawhub@latest install book-deep-reader --registry https://cn.longxiaskill.com
技能文档
Produce teaching-quality book notes: accurate on the original, structured for quick learning, detailed enough to explain to others.
Workflow
Phase 1: Identify the Book
Given a book (title, author, ISBN, or URL), gather foundational info:
- Search for the book using
mimo_web_search— find Douban/Goodreads pages, publisher descriptions, table of contents, reviews - Fetch key pages using
web_fetch— Douban book page (ISBN lookup), publisher pages, review articles - Identify: book structure (chapters), core framework/model, key cases, author's background
- For companion/sequel books: also fetch the original book's wiki/summary to cross-reference (e.g., Reinventing Organizations Wiki for the practice guide)
Phase 2: Deep Research
Gather rich source material from authoritative references:
- Official wiki/community sites (if they exist) — e.g., reinventingorganizationswiki.com
- Detailed reviews and analysis articles — not just "good book!" but substantive breakdowns
- Author's talks/interviews — YouTube transcripts, keynote summaries
- Case study databases — for books with organizational/practice cases
Search strategy (use mimo_web_search):
"" 目录 章节 内容(for Chinese books)"" summary chapters key concepts"author name" "" detailed explanationsite:douban.comorsite:goodreads.com