dsh-daoing-memory
Run the following command in DeepSeek Harness:
dsh plugin install daoing/dsh-daoing-memory
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install daoing/dsh-daoing-memory in your terminal, or install from the source repository at https://github.com/daoing/dsh-daoing-memory, then restart DSH to activate the memory tools.
About this plugin
Every DSH session starts from zero, and most memory bolt-ons make it worse: they either dump raw conversations into a vector store (burying signal in noise) or let the model write whatever it wants into a key-value blob (one hallucination poisons every future session). dsh-daoing-memory takes a stricter stance: memory must be earned. An experience is born as a low-trust candidate and is promoted only after it has been corroborated by real, repeated use — no single write gets a free pass to high trust.
Organized around four verbs — Generate, Use, Revise, Record — the plugin keeps semantic memory (durable facts and open concerns about the user) strictly separate from experiential memory (reusable how-to knowledge with a lifecycle). Diary entries, structured extraction, experience trust escalation, keyword-and-situation recall, an append-only audit ledger, periodic consolidation, and a compact profile snapshot auto-injected into the system prompt form a complete closed loop. A browser workbench lets you read, correct, promote, or delete any entry at any time, so memory is a shared, human-curated artifact rather than a black box.
Built for developers who want their DSH agent to genuinely know the user across sessions, accumulate verifiable experience, and treat memory quality as a first-class concern. If you are tired of vector-stuffing and free-form KV writing, and want an auditable, reversible, human-in-the-loop memory foundation, this is the plugin to try.
Use Cases
- Retain a user profile, facts, and open concerns across sessions
- Accumulate, verify, and revise reusable experiences over multiple conversations
- Inspect and curate memory entries via a browser workbench
Best For
- Developers who want their DSH agent to truly know the user across sessions
- Teams that require auditable and reversible memory guarantees
- Engineers who are done with vector-dump and free-form KV memory
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.