dsh-evolve
Run the following command in DeepSeek Harness:
dsh plugin install chenzheshushi-commits/dsh-evolve
Paste the following prompt into your AI chat to install this plugin:
In DeepSeek Harness, install this plugin using the command 'dsh plugin install chenzheshushi-commits/dsh-evolve'. The source code is at https://github.com/chenzheshushi-commits/dsh-evolve.
About this plugin
Intelligent agents often forget everything between sessions, leading to redundant learning, inefficiency, and accumulated memory that turns into a noise pile. The dsh-evolve plugin addresses this core issue by providing a self-evolving memory and skill lifecycle for DeepSeek Harness. It offers persistent cross-session memory with structured records, crystallizes repeated procedures into reusable skills, and introduces tiered approval, reinforcement learning, and anti-bloat convergence mechanisms to ensure memory never grows chaotic. From facts to preferences and decisions, memories are stored durably, recalled intelligently, and injected automatically when relevant, allowing agents to build real experience rather than losing context.
The plugin's core capabilities include a cross-session memory system with zero-token deterministic recall; skill crystallization and versioning with a full lifecycle from active to archived; and anti-bloat convergence that detects and merges duplicate skills or prunes low-value memories to prevent system bloat. All operations follow design principles of never breaking the harness or deleting user assets, with backups and rollbacks for safety. Background review and maintenance tools further optimize memory quality, while a human-facing prune page makes control intuitive.
dsh-evolve is ideal for developers, AI researchers, or anyone using DeepSeek Harness who needs long-term memory management. If you aim to build agents that continuously learn and improve without accumulating noise, this plugin provides a flexible framework. It ships blank with no preloaded data, learning is localized for privacy, and it allows deep customization. Whether for personal assistants or complex automation, it helps agents evolve from experience rather than being limited to single conversations.
Use Cases
- Intelligent agents forget context between sessions and need persistent memory support.
- When repeating operations, the desire to automatically crystallize reusable skills.
- Memory accumulates into noise, requiring intelligent cleanup and optimization management.
Best For
- Developers using DeepSeek Harness.
- AI researchers or intelligent agent builders.
- Users who need to manage long-term memory and skills.
Related Plugins
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.
An opinionated, zero-infrastructure file-based memory protocol: human-readable markdown files plus a bounded always-loaded index, curation discipline, and no database or embeddings.