Core Features

This chapter dives deep into DCP's three core capabilities: auto-pruning strategies, LLM-driven tools, and Slash commands. After mastering these features, you can fully leverage DCP's token optimization potential.

Chapter Contents

Auto-Pruning Strategies

Deep dive into how DCP's three automatic strategies work: deduplication strategy, overwrite strategy, and error clearing strategy. Learn the trigger conditions and use cases for each strategy.

LLM-Driven Pruning Tools

Understand how AI autonomously calls discard and extract tools to optimize context. This is DCP's most intelligent feature, allowing AI to actively participate in context management.

Slash Command Usage

Master all DCP command usage: /dcp context to view context status, /dcp stats to view statistics, /dcp sweep to manually trigger pruning.

Learning Path

We recommend studying the contents of this chapter in the following order:

Auto-Pruning Strategies → LLM-Driven Tools → Slash Commands
          ↓                    ↓                  ↓
     Understand Principles   Master Smart Pruning   Learn Monitoring & Debugging

1. First learn auto-pruning strategies: This is the foundation of DCP, understand the working principles of the three strategies
2. Then learn LLM-driven tools: After understanding automatic strategies, learn more advanced AI active pruning capabilities
3. Finally learn Slash commands: Master monitoring and manual control methods for debugging and optimization

Prerequisites

Before studying this chapter, ensure you have completed:

• Installation & Quick Start - Complete DCP plugin installation
• Configuration Guide - Understand basic concepts of the configuration system

Next Steps

After completing this chapter, you can continue exploring:

• Protection Mechanisms - Learn how to protect critical content from being mistakenly pruned
• State Persistence - Understand how DCP preserves state across sessions
• Prompt Caching Impact - Understand the trade-offs between DCP and Prompt Caching