dsh-memory-panel
Run the following command in DeepSeek Harness:
dsh plugin install x2it/dsh-memory-panel
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install x2it/dsh-memory-panel in the DeepSeek Harness terminal; the source code is hosted at https://github.com/x2it/dsh-memory-panel. After installing, restart dsh web to see the Memory panel in the sidebar.
About this plugin
Every new session in DeepSeek Harness starts from a blank slate. Preferences you stated last week, project constraints, half-finished task context - all gone the moment the conversation closes. dsh-memory-panel adds a durable, cross-session memory layer: the agent gains five tools (memory_save, memory_recall, memory_list, memory_delete, memory_clear) to structure information by key and category, while a bilingual sidebar panel lets you inspect at a glance exactly what the agent has stored about you. All data lives in a single local JSON file at ~/.dsh/memory.json, written atomically with concurrent-write serialization, automatic corruption backup, and hard caps on entry count and per-entry length so the store stays lean and recoverable.
If you want dsh to behave less like a stateless endpoint and more like a long-term collaborator that remembers your stack, your team conventions, and the projects in flight, this plugin is the lightest local-first answer - no vector database, no external service, just a file you can read, audit, and reset at any time. It suits individual developers and small teams who value data sovereignty and want context continuity across sessions without adding infrastructure.
The growth controls (500-entry cap, 4000-char per-entry limit, 30-result recall ceiling) keep the store predictable even as it ages, and the corruption-recovery path silently backs up a broken memory.json before starting fresh, so you never face a silent data-loss surprise.
Use Cases
- Preserve project context and team conventions across sessions instead of repeating the stack and conventions every time.
- Audit and review the agent's stored memory entries at a glance via the sidebar panel, ensuring no unintended data lingers.
- Manage personal preferences and task state in a single local JSON file without a vector database or external cloud service.
Best For
- Individual developers using DeepSeek Harness for daily work who want cross-session agent memory.
- Privacy-conscious users who prefer local storage over cloud-based services and value data sovereignty.
- Small teams that need consistent project conventions and want to reduce repetitive communication overhead.
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.