长文档智能总结

        @OpenClaw I need you to summarize a long document.

Context: I am using openclaw.

Step 1 — Environment check
- Check whether @modelcontextprotocol/server-fetch is installed.
- If missing, ask whether to install it.

Step 2 — Get the document
Ask me to choose one:
1) Provide a public URL
2) Paste the content directly
3) I have a local file (tell me to upload it to a public URL first)

If I provide a URL, use Fetch MCP to retrieve it.

Step 3 — Analyze
1) Identify document type and audience
2) Outline the structure (sections)
3) Extract 3–5 core points, key facts/data, conclusions

Step 4 — Write the summary (markdown)
```markdown
# 📄 Document Summary

## Document Info
- Title: [Title]
- Type: [Type]
- Length: ~[Words]
- Est. reading time: [Minutes]

## 🎯 One-sentence takeaway
[One sentence]

## 📌 Key points
1. **[Point 1]** — [2–3 sentences]
2. **[Point 2]** — [2–3 sentences]
3. **[Point 3]** — [2–3 sentences]

## 📊 Key data
- [Data 1]
- [Data 2]

## 💡 Conclusions
[Conclusions + recommendations]
```

Step 5 — Next action
Ask whether I want:
1) per-section deep summary
2) extract only a specific topic
3) translate the summary
4) done

Safety guarantees
- Only read what I provide (public URL / pasted text)
- Do not store or upload the content elsewhere

整个流程请用简体中文与我交互。