dsh-agent-memory
Run the following command in DeepSeek Harness:
dsh plugin install xingyingyuzhui/dsh-agent-memory
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install xingyingyuzhui/dsh-agent-memory in DeepSeek Harness to install this plugin; the source repository is https://github.com/xingyingyuzhui/dsh-agent-memory. We recommend installing dsh-agent-registry and dsh-agent-identity first, then restarting dsh web after installation.
About this plugin
Claw Agents have no natural long-term memory across turns. The moment official compaction fires, the context and decisions accumulated over a session vanish with the window. dsh-agent-memory gives every Agent its own vault: two persistent documents (USER and MEMORY) plus a dated diary, so the Agent genuinely remembers between turns.
Three events drive the core loop. After each turn, a post-round review (toggleable) distills key information into the vault. When the official session/event emits compaction/start, the current working content is flushed to the vault immediately to survive compression. The memory tool attaches only to Claw sessions, never intercepts other tools, and workspace or official sessions remain free of Claw memory. Writes can be free-form or approval-gated; pending items appear in the Agent Memory tab.
Built for dsh users running multiple Claw Agents who want isolated per-Agent long-term memory with zero cross-contamination. All data lives locally under the .dsh directory; uninstalling the plugin leaves files on disk but stops injection and review.
Use Cases
- Retaining cross-session context across multi-turn conversations
- Auto-archiving key info before official compaction
- Maintaining isolated per-Agent memory spaces
Best For
- dsh users running multiple Claw Agents
- Teams needing long-term Agent memory without cross-contamination
- Privacy-sensitive users who prefer local-only data storage
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.