閱讀助手
Analysis
This is a coherent EPUB reading assistant, but users should notice that it installs Python packages, stores book text and summaries locally, and can optionally sync or run reminders.
Findings (5)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
pip3 install ebooklib beautifulsoup4 lxml
The skill directs installation of third-party Python packages without pinned versions or an install spec. This is purpose-aligned for EPUB parsing, but it changes the local Python environment.
將章節文字交給 LLM(即你自己),以下方結構產生摘要
The workflow passes arbitrary EPUB chapter text into the model for summarization. That is central to the skill, but imported book text should be treated as content to summarize, not instructions to obey.
"schedule": "0 8 * * *", "prompt": "查看我的閱讀進度,如果有正在讀的書,幫我摘要下一章。"
The documentation shows an optional cron job that would repeatedly trigger the skill to summarize the next chapter and push it to a communication platform. It is disclosed and user-configured, but it is persistent autonomous behavior.
Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.
如果使用者有連接 Notion,可同步書籍資訊與摘要到 Notion 資料庫
The skill can use a connected Notion account to create or update a reading database. This is optional and disclosed, but it relies on external account authority not listed as a primary credential.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
書庫根目錄:`~/.openclaw/workspace/reading-library/` ... 閱讀進度檔:`~/.openclaw/workspace/reading-library/progress.json`
The skill persistently stores imported book chapters, reading progress, and generated summaries. This is disclosed and purpose-aligned, but it creates reusable local reading data.
