Quote Extractor

Skill Positioning

• ❌ No admission audit (does not judge whether it qualifies for storage)
• ❌ No content processing (does not extract cognitive frameworks)
• ✅ Focus on golden sentence extraction
• ✅ Suitable for quickly accumulating writing materials

When to Use

1. The user explicitly requests "extract golden sentences", "excerpt good sentences", "save these expressions"
2. The user says "This sentence is well-written", "This metaphor is brilliant" while reading an article
3. The user enters the shortcut command:

   /extract-quotes

   or

   /extract-quotes [file path]

• If the user needs to organize external materials and store them in the library → Use

  /kn

  (knowledge_auditor)
• If the user wants to learn the article structure → Use

  /analyze-writing

  (writing_analyzer)

Core Workflow

Step 1: Read Article Content

1. Obtain Input:
   • If the user provides a file path: Read directly
   • If the user pastes content: Use the pasted text
   • If the user provides neither: Ask the user to provide a file path or paste content
2. Confirm Extraction Criteria:
   • Ask the user: "What type of golden sentences do you want to extract?"
     • Comparative Golden Sentences (A vs B type expressions)
     • Metaphor & Analogy (vivid metaphors for abstract concepts)
     • Counterintuitive Viewpoints (expressions that subvert conventional cognition)
     • Call to Action (inspiring concluding sentences)
     • All Types (default)

Step 2: Identify Golden Sentences

Criterion 1: Uniqueness

• ✅ Not a common expression (e.g., "We must work hard")
• ✅ Has a unique perspective or metaphor
• ✅ Stands out and impresses

Criterion 2: Quotability

• ✅ Can stand alone and be understood without the original context
• ✅ Conveys a complete meaning without context
• ✅ Suitable for direct use in one's own articles

Criterion 3: Emotional Impact

• ✅ Has impact or resonance
• ✅ Makes readers pause and think
• ✅ Has a vivid imagery or rhythmic flow

Golden Sentence Type List

Step 3: Extract and Classify

1. Sentence-by-Sentence Scan:
   • Read the full text, identify sentences that meet the criteria
   • Retain the original expression (no rewriting)
   • Mark the paragraph position (for easy context retrieval)
2. Classify and Organize:
   markdown

   ## Comparative Golden Sentences
   - "Exams require existing answers and correctness; the market requires usefulness and problem-solving"
     [Source: Paragraph 3]
   
   ## Metaphor & Analogy
   - "Treat your current girlfriend the way you treated your ex, and you can imagine the result"
     [Source: Opening]
   
   ## Call to Action
   - "Don't learn first then act, act first then learn"
     [Source: Paragraph 5]

3. Quality Screening:
   • Keep only 3-10 sentences (avoid "hoarding")
   • Prioritize the most impactful ones
   • If more than 10 sentences are identified, ask the user "Do you want to save all of them?"

Step 4: Generate Material Notes

---
type: Material
class: Golden Sentence Library
level: B
summary: Golden sentence materials related to [Topic]
review_freq: On demand
source: [Article Title / Source]
tags: [Content Creation, Golden Sentence Materials, [Topic Keywords]]
create_date: [Today's Date]
aliases: [[Topic] Golden Sentences, [Topic] Copywriting]
---

## Usage Instructions

**Topic**: [One-sentence description of the topic of these golden sentences, e.g., "Workplace Mindset Shift", "Time Management", "Knowledge Management"]
**Applicable Scenarios**: [When these golden sentences can be quoted]
**Notes**: Indicate the source when quoting (can mark "Rewritten from XXX")

---

## Golden Sentence List

### Comparative Category
> [Golden Sentence 1]
- **Applicable Scenario**: [When to use]
- **Source Paragraph**: [Original Paragraph X]

> [Golden Sentence 2]
- **Applicable Scenario**: [When to use]
- **Source Paragraph**: [Original Paragraph X]

### Metaphor & Analogy Category
> [Golden Sentence 3]
- **Applicable Scenario**: [When to use]
- **Source Paragraph**: [Original Paragraph X]

### Call to Action Category
> [Golden Sentence 4]
- **Applicable Scenario**: [When to use]
- **Source Paragraph**: [Original Paragraph X]

---

## Extended Thinking

**Rewriting Directions**: [How to extend your own expressions based on these golden sentences]
**Combined Usage**: [Which sentences can be used together]

---

## Reference Sources
- Original Text: [Article Title]
- Extraction Date: [Today's Date]

Step 5: Inventory Retrieval and Deduplication

1. Duplicate Check: Retrieve in the directory

   E:\OBData\ObsidianDatas\3通用技能\内容创作\Attachment(素材库)

   :
   • Keywords: topic keywords, similar expressions
   • Check if there are existing golden sentence notes on the same topic
2. Proposals:
   • Scenario A (Append Mode): If a golden sentence note on the same topic is found, suggest appending to the existing note (avoid fragmentation)
   • Scenario B (New Note Mode): If no inventory is found or the topic is completely different, create a new note
3. Archiving Location:

   E:\OBData\ObsidianDatas\3通用技能\内容创作\Attachment(素材库)\[Topic] Golden Sentence Materials.md

Step 6: Execution and Delivery

1. Ask for User Confirmation:
   • Display the extracted golden sentence list
   • Ask: "Do you want to keep all these golden sentences? Do you need to delete or add any?"
2. Execute Writing:
   • After user confirmation, physically write to the file
   • Return the file path
3. Follow-up Suggestions:
   • Remind the user to delete the source file (if extracted from an external article)
   • Suggest reviewing the golden sentence library regularly (once a month)

Interaction Style (Tone & Rules)

1. Restraint Principle: Strictly control the quantity (3-10 sentences) to avoid "material hoarding"
2. Original Flavor: Retain the original expression, no rewriting (unless requested by the user)
3. Scenario-based: Each golden sentence must be marked with "Applicable Scenario"
4. No Empty Talk: Must use the

   Write

   tool to physically write to the file

Calling Examples

Example 1: Specify File Path

/extract-quotes E:\OBData\ObsidianDatas\0收集箱日清\考察思维 vs 市场思维.md

Example 2: Paste Content

/extract-quotes

[User pastes article content]

Example 3: Real-time Extraction

User: This sentence is well-written: "Your energy is not a light bulb, but a laser"
Assistant: Sure, I'll help you extract this golden sentence. May I ask which article this sentence is from? What is the topic?

Golden Sentence Quality Self-Check List

• Uniqueness: Is it not a common expression?
• Quotability: Can it be understood without the original context?
• Emotional Impact: Does it have impact or resonance?
• Applicability: Can it be used in writing next week?

Cooperation with Other Skills

• knowledge_auditor (

  /kn

  ): If the user needs a complete audit + storage process, transfer to it
• writing_analyzer (

  /analyze-writing

  ): If the user says "I want to learn the writing structure of this article" after extracting golden sentences, recommend using the writing analyzer
• thought-mining (

  /thought-mining

  ): If the user wants to write a new article based on the golden sentences, recommend using the thought mining assistant

Notes

1. Copyright Awareness: Remind the user to indicate the source when quoting
2. Avoid Hoarding: If the user frequently extracts but never uses, issue a warning: "You have accumulated XX golden sentences, but haven't used any in the last 3 months. Use them up before extracting more."
3. Regular Cleanup: Remind the user every quarter to clean up golden sentences that "have never been quoted"