dsh-plugin-thread
Run the following command in DeepSeek Harness:
dsh plugin install zhaoyuntao-wl/dsh-plugin-thread
Paste the following prompt into your AI chat to install this plugin:
To install this plugin in DeepSeek Harness, please run: `dsh plugin install zhaoyuntao-wl/dsh-plugin-thread`. Source: https://github.com/zhaoyuntao-wl/dsh-plugin-thread
About this plugin
When building complex coding agents, losing session memory or obscuring decision history are often the biggest pain points. dsh-plugin-thread aims to solve this by providing lossless event stream capture and structured lineage management within SQLite, ensuring every interaction is fully recorded and traceable. Leveraging deterministic log design, it supports first-turn anchoring, re-anchoring after compaction, and cross-agent state deltas, perfectly avoiding memory failure modes like the "compaction cliff."
The plugin not only offers native query tools similar to a file system (ls/cat/grep) but also introduces a "behavior contract" skill that allows models to be aware of their own memory without explicit checks. It registers decisions, preferences, and assets with clear commands for management, complemented by a todo list and candidate queue to provide a robust yet flexible memory layer. This enables developers to build agent systems that maintain state consistency, logical clarity, and auditability during long-running operations.
This plugin is particularly suitable for developers working within the DeepSeek Harness environment, especially those who need to handle long-term projects, complex logical chains, and high-concurrency sessions. It runs completely locally without backend dependencies, ensuring data privacy and simplifying deployment. Whether for pursuing code quality or team collaboration efficiency, dsh-plugin-thread serves as an indispensable memory hub for agent development.
Use Cases
- State memory and recovery for long-term coding projects
- Decision lineage tracking in complex code generation tasks
- Localized data persistence in agent workflows
Best For
- Developers building agents using DeepSeek Harness
- AI engineers requiring reliable memory systems
- Teams focused on long-term code generation projects
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.