Personal Material Library Search
Quickly search through Peanut's real experiences, perspectives, and cases to add a human touch and credibility to content creation.
When to Use This Skill
Automatically load this skill when the following scenarios are detected:
- Users say "Find my previous experiences", "Are there relevant cases"
- Articles/scripts need support from real cases
- Content is too "AI-like" and needs to add a human touch
- Real evaluations of tools/products are needed
- Personal experiences are required as opening remarks
Core Features
- Quick Search: Search keywords across all Jike posts (1800+ entries)
- Smart Matching: Automatically match relevant materials based on topics
- Rewrite Suggestions: How to expand short posts into long-form article logic
- Authenticity Guarantee: All materials are Peanut's real expressions
Material Library Overview
Data Sources
- All Jike Posts.csv: 1800+ real posts (May 2025 - October 2025)
- My AI Exploration Station Posts.csv: AI-related posts
- Refined Thematic Materials: High-quality materials categorized by topic
Main Topic Distribution
- AI Programming Tools: 504 entries (Cursor, Claude Code, GitHub Copilot, etc.)
- Product Development: 200 entries (HarmonyOS development, Kitten Fill Light, automated operations)
- AI Model Evaluation: 150 entries (GPT, Claude, DeepSeek, etc.)
- Self-media Operation: 100 entries (growth strategies, content creation)
- Technical Learning: 80 entries (programming experience, pitfall records)
Search Methods
Method A: Directly Search Raw Data (Recommended Priority) ⭐
Applicable Scenarios:
- New topics, unorganized themes
- Need the most original real expressions
- Current affairs comments, hot topics
Operation Steps:
- Use Grep Tool for Search
pattern: "keyword1|keyword2|keyword3"
path: /Users/alchain/Documents/Writing/Writing References/Personal Material Library/All Jike Posts.csv
output_mode: content
-n: true
-i: true # Ignore case
- Extract Core Perspectives
From search results, find:
- Peanut's real complaints
- Specific observations and attitudes
- Actual usage experiences
- Data and achievements
- Rewrite into Article Materials
- ✅ Expand into long-form logic (short posts → complete paragraphs)
- ✅ Maintain real attitudes and complaining tone
- ✅ Add background explanations and details
- ❌ Do not directly copy the original text
Search Examples:
# Write articles related to Amap Street Sweeping Rank
pattern: "Amap|Street Sweeping Rank"
# Write AI programming tool comparisons
pattern: "Cursor|Claude Code|Copilot|GitHub"
# Write product growth strategies
pattern: "growth|conversion|users|retention"
# Write technical learning experiences
pattern: "learning|mastery|understanding|pitfalls"
# Write self-media operation
pattern: "Official Account|videos|Xiaohongshu|play volume|reading volume"
Advantages:
- Quickly find Peanut's real perspectives
- Obtain the most original expressions and emotions
- Suitable for all topics, not limited by organized materials
- Can find the latest posts (updated to October 2025)
Method B: View Refined Materials
Applicable Scenarios:
- Common topics (AI programming tools, product development, etc.) have systematic organization
- Need complete experience summaries
- Looking for systematic perspectives
Operation Steps:
- Open the Master Index
/Writing References/Personal Material Library/README.md
- Find Topics
Open
_Jike Posts Thematic Index.md
and find relevant topics:
- AI programming tools? → 504 materials
- Product development? → 200 materials
- AI model evaluation? → 150 materials
-
Select Materials
View refined files (e.g.,
AI Programming Tool Usage Experience.md
)
-
Rewrite and Use
Refer to examples in
_Jike Posts Usage Specifications.md
:
- ✅ Expand into long-form logic
- ✅ Maintain real attitudes
- ❌ Do not directly copy
List of Refined Topics:
AI Programming Tool Usage Experience.md
Product Development Experience Summary.md
Self-media Operation Practice.md
Technical Learning Methods.md
Typical Usage Scenarios
Scenario 1: Article Opening (Introduce with Real Experience)
Requirement: Write In-Depth Evaluation of Claude Code, need an opening
Search:
Found Material:
"Claude Code's biggest advantage is context understanding. When modifying a feature, it knows which files to touch."
Rewrite Example:
I've been using Claude Code for two weeks. The most obvious change is that before modifying a feature, I had to manually check 5-6 files, but now Claude Code can find all of them at once. This capability is crucial, especially when taking over old projects.
Today let's talk about whether Claude Code is worth using.
Scenario 2: Perspective Support (Real Evaluation)
Requirement: Write Guide to Choosing AI Programming Tools, need real comparisons
Search:
pattern: "Cursor|Claude Code|comparison"
Found Material:
"Cursor is suitable for small projects, Claude Code is suitable for large projects. Price is a major drawback."
Rewrite Example:
I've used both Cursor and Claude Code for 3 months. Here's my conclusion:
- Cursor: Suitable for personal projects, small teams. $20 per month, high cost-effectiveness.
- Claude Code: Suitable for large projects, complex refactoring. $100+ per month, suitable for company use.
If you're an individual developer, start with Cursor. If your company covers the cost, go for Claude Code.
Scenario 3: Case Display (Real Project)
Requirement: Write Introduction to HarmonyOS Development, need real project cases
Search:
Found Material:
"The biggest pitfall of HarmonyOS development is incomplete documentation. Official examples don't run."
Rewrite Example:
I previously used Cursor to develop a HarmonyOS app and encountered many pitfalls. The biggest problem is incomplete documentation — 30% of official examples don't run, and error messages are unclear.
Later I found that the fastest way is to directly look at the source code. HarmonyOS component source code is written quite clearly, which is more reliable than the documentation.
Scenario 4: Closing Reflection (Personal Insight)
Requirement: Write The Value of Programmers in the AI Era, need a concluding升华
Search:
pattern: "AI|replace|programmer|value"
Found Material:
"AI can write for me, but it can't think for me. The real value is judgment."
Rewrite Example:
AI can write code for me, but it can't figure out why I need to write this code.
Technology will become stronger and stronger, but judgment will always be human — choosing which tech stack to use, what product to build, who to serve, how to grow. This is something AI can't learn.
The value of a programmer has never been about the speed of typing code, but about figuring out what code to type.
Rewriting Specifications
✅ Correct Rewriting
- Expand Background: Turn short posts into complete paragraphs, add time, scenarios, reasons
- Maintain Attitude: Don't "polish" real complaints and clear views into official jargon
- Add Details: Specific numbers, timelines, operation steps
- Natural Transition: Integrate materials into articles without being abrupt
❌ Incorrect Rewriting
- Direct Copy: Paste short posts verbatim
- Over-Polishing: Change "encountered pitfalls" to "learning opportunities"
- Lose Personality: Change complaining tone to neutral and objective
- Fabricate Details: Don't add data not present in the original material
Rewriting Example Comparison
Original Material (Jike post):
"Amap Street Sweeping Rank forced pop-up during National Day, so annoying. The product manager has a screw loose."
❌ Incorrect Rewriting (Over-Polishing):
"Amap still has room for improvement in user experience, especially the design of pop-up functions needs to be more user-friendly."
✅ Correct Rewriting (Maintain attitude, expand details):
"When I opened Amap during National Day, a Street Sweeping Rank activity page popped up forcefully, and I couldn't close it without clicking 'Participate'. I just wanted to navigate, why do I have to be forced to join the activity?
This design logic is confusing. What's the core demand of users opening navigation? It's to start navigation quickly, not to join the Street Sweeping Rank. Forced pop-ups only make people annoyed, and they don't help with activity participation rates — will people who are forced to click 'Participate' actually go street sweeping?
The core of product design is to respect user intent, not to push KPIs forcibly."
Notes
1. Authenticity Principle
- ✅ All materials are real, do not fabricate or exaggerate
- ✅ Data, time, tool names must be accurate
- ✅ Do not pass off others' experiences as Peanut's
2. Timeliness Principle
- ✅ Pay attention to the time span (May 2025 - October 2025)
- ✅ Avoid using outdated information (e.g., "GPT-3.5 is mainstream")
- ✅ Verify if product functions are still applicable when involving them
3. Rewriting Principle
- ✅ Must rewrite into long-form logic, do not directly copy
- ✅ Maintain Peanut's real attitude and complaining tone
- ✅ Add background, details, thinking process
- ❌ Do not "polish" into official jargon or neutral objective language
4. Usage Frequency
- ✅ Use at least 1-2 real materials per article
- ✅ Opening, cases, closing are the best positions to use
- ❌ Do not over-rely on materials, maintain originality
Frequently Asked Questions
Q1: What if I can't find relevant materials?
A:
- Try adjusting keywords (synonyms, related concepts)
- Expand search scope (search superordinate concepts)
- Check the thematic index to see if there are relevant categories
- If there are really no materials, you can do without them and supplement through search and analysis
Q2: What if the materials don't fully match the article topic?
A:
- Extract core perspectives and attitudes instead of specific content
- Introduce by analogy ("This reminds me of when...")
- Use as counterexamples or for comparison
Q3: How to judge if materials are outdated?
A:
- Check the time tag (release time is included in CSV)
- When involving product functions, search to verify the latest status
- If in doubt, mark "At that time..." or "In X month..."
Q4: Must I use personal materials for every article?
A:
- Not mandatory, but strongly recommended
- Real materials are the most effective way to reduce AI-like feel
- Use at least 1-2 materials in the opening or case section
Technical Implementation
Grep Search Example
bash
# Search keywords (supports regular expressions)
grep -i "Cursor\\|Claude Code" /Users/alchain/Documents/Writing/Writing References/Personal Material Library/All Jike Posts.csv
# Search and display line numbers
grep -in "growth\\|conversion" /Users/alchain/Documents/Writing/Writing References/Personal Material Library/All Jike Posts.csv
# Search and display context (3 lines before and after)
grep -i -C 3 "HarmonyOS" /Users/alchain/Documents/Writing/Writing References/Personal Material Library/All Jike Posts.csv
Using Grep Tool in Claude Code
Grep tool parameters:
pattern: "keyword1|keyword2"
path: /Users/alchain/Documents/Writing/Writing References/Personal Material Library/All Jike Posts.csv
output_mode: content
-n: true
-i: true
-C: 2 # Display context
Reference Resources
/Writing References/Personal Material Library/README.md
- Master Index (Start here)
/Writing References/Personal Material Library/All Jike Posts.csv
- Raw Data (1800+ entries)
/Writing References/Personal Material Library/_Jike Posts Thematic Index.md
- Thematic Classification
/Writing References/Personal Material Library/_Jike Posts Usage Specifications.md
- Rewriting Specifications and Examples
Success Cases
- Why the Best Content is Useless: Used MrBeast case materials to increase authenticity
- A Team Burning Billions of Tokens a Month: How They Use Claude Code: Used multiple Claude Code usage experiences, high credibility
Last Updated: 2025-11-07
Applicable Projects: Official Account Writing, Video Scripts
Maintainer: Peanut
Produced by Uncle Hua | AI Native Coder · Independent Developer
Official Account "Uncle Hua" | 300,000+ followers | AI Tools and Efficiency Improvement
Masterpieces: Kitten Fill Light (Top 1 in AppStore Paid Rank) · Mastering DeepSeek in One Book
",