AI-Powered Knowledge Base Cleanup Skill
A 5-step progressive cleanup AI Skill for intelligent knowledge base organization. Built for AI agents (Reasonix, Claude Code, AutoGen, LangChain) and developers.
| Round | Operation | Description |
|---|---|---|
| R1 | MD5 Exact Duplicates | Byte-level duplicate file removal |
| R2 | Filename Similarity | Version chain detection (V1→V2→V3...) |
| R3 | Aggressive Normalization | Automatic implicit version discovery |
| R4 | Archive Detection | Safe removal of extracted archives |
| R5 | Directory Reorganization | Project/Management two-tier structure |
- 🔍 Intelligent Detection: Semantic analysis for file versioning patterns
- 🛡️ Safe Operations: Read-only source, backup before every modification
- 📊 Detailed Reporting: Comprehensive cleanup reports for each round
- 🔄 Rollback Support: Full rollback capability for each round
- 🤖 AI Agent Integration: Native Reasonix Skill support
- 🚀 Cross-Platform: Works on Windows, macOS, and Linux
git clone https://github.com/CS-Faith/knowledge-cleanup.git
cd knowledge-cleanup
python run_cleanup.py <source_directory> <target_directory>- 9,676 files → 7,579 files (-21.7% cleanup rate)
- ~2GB space freed in typical cleanup
- Safe execution with user confirmation at each step
- Reasonix: Native Skill support with
/skill knowledge-cleanup - Claude Code: Direct execution support
- AutoGen: Tool integration for AI agents
- LangChain: Compatible as a LangChain tool
- Automatic backups before every modification
- Full rollback capability for each round
- Source directory is read-only - never modified
- All operations happen in target directory only
MIT License - see the LICENSE file for details.
- GitHub: CS-Faith
- Repository: knowledge-cleanup
| 项目 | 描述 | 链接 |
|---|---|---|
| reasonix-portakit | 便携工具箱 | CS-Faith/reasonix-portakit |
| reasonix-migration-assistant | 配置迁移升级助手 | CS-Faith/reasonix-migration-assistant |
| llm-wiki-pipeline | 知识库构建流水线 | CS-Faith/llm-wiki-pipeline |