docs: Update file paths and structure in documentation

- Consolidate AGENTS.md and CLAUDE.md under libs/external/.
- Update README.md and GEMINI.md to reflect the new paths for AGENTS.md, CLAUDE.md, and prompts-library.
- Correct internal references in libs/external/AGENTS.md and libs/external/CLAUDE.md.

libs/external/AGENTS.md

@@ -5,4 +5,4 @@ https://github.com/yusufkaraaslan/Skill_Seekers
 
 https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools
 
-我的nvim配置（来自https://github.com/LazyVim/LazyVim），./libs/common/utils/my-nvim
+我的nvim配置（来自https://github.com/LazyVim/LazyVim），./libs/external/my-nvim

libs/external/CLAUDE.md

@@ -0,0 +1,52 @@
+# CLAUDE.md
+
+This file provides guidance to Claude series models when working with code in this repository.
+
+## Repository Overview
+
+This is the **Vibe Coding CN** repository, a workflow, toolset, and knowledge base for advanced AI-assisted programming. The project's core assets are its extensive `prompts` and `skills` libraries.
+
+## Key Commands
+
+### Prompt Library Management
+```bash
+# Enter the library directory
+cd libs/external/prompts-library
+
+# Run the interactive conversion tool
+python3 main.py
+```
+
+### Development & Maintenance
+```bash
+# Lint all markdown files in the repository
+make lint
+
+# Create a full project backup (respects .gitignore)
+bash backups/一键备份.sh
+```
+
+## Architecture & Structure
+
+### Core Directories
+- **`prompts/`**: The core asset. A massive, well-organized library of prompts.
+  - `coding_prompts/`, `system_prompts/`, `user_prompts/`
+- **`skills/`**: A modular library of skills for the AI, providing domain-specific knowledge for various tools like `ccxt`, `postgresql`, `telegram-dev`, etc.
+- **`documents/`**: The project's knowledge base, containing methodology, principles, and guides.
+- **`libs/external/prompts-library/`**: A Python-based tool for converting prompts between Excel and Markdown formats.
+- **`backups/`**: Scripts for project backups.
+- **`libs/`**: Skeleton for shared Python library code.
+
+### Key Technical Details
+1.  **Prompt Organization**: Prompts use a `(row,col)_` prefix for categorization.
+2.  **Conversion Tool**: The `prompts-library` uses Python with `pandas` and `openpyxl`.
+3.  **Documentation Standard**: User-facing documentation is in Chinese. Code, file names, and structure are in English.
+4.  **Skills**: The `skills` directory provides context and knowledge for specific tools and domains, each with its own `SKILL.md`.
+
+## Development Workflow
+
+When modifying this repository:
+1.  Follow the existing prompt and skill categorization systems.
+2.  Use the `prompts-library` tool to maintain consistency when updating prompts.
+3.  Run `make lint` after changing any Markdown files.
+4.  Run a backup with `bash backups/一键备份.sh` before any major refactoring.