MIT-licensed native DSH memory plugin: three tiers (Runtime / Documents / Memory Spaces) with nine pluggable long-term providers and a supervised, Host-controlled write workflow. Default engine is local-first Mnemon.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-08-23
License
MIT
01What can it help you accomplish?
Give DeepSeek Harness one memory control plane across runtime context, project documents, and cross-session durable facts
Compact per-turn Runtime Memory (USER.md / MEMORY.md), searchable Project Documents, and bounded recall from Memory Spaces — without forcing every kind of knowledge into one database
DSH users and agents who need persistent, cross-session, multi-agent memory with pluggable long-term providers instead of a single flat store
02How to install into DeepSeek Harness
Prerequisites
- DeepSeek Harness (DSH) — the release is verified against DSH 0.1.1-rc.2, whose complete profiles require Node.js ^22.19.0 || >=24.0.0
- Mnemon Native (the default, prioritized engine): install via Homebrew on macOS, or Go
Installation steps
- 01
Install Mnemon Native: `brew install --cask mnemon-dev/tap/mnemon` (macOS) or `go install github.com/mnemon-dev/mnemon@latest`, then `mnemon --version`
$ brew install --cask mnemon-dev/tap/mnemon
- 02
Install DSH: `npm install -g @deepseek-ai/dsh@0.1.1-rc.2`, then `dsh --version`
$ npm install -g @deepseek-ai/dsh@0.1.1-rc.2
- 03
Add the plugin: `dsh plugin --profile web add dsh-mnemon`, then `dsh --profile web`
$ dsh plugin --profile web add dsh-mnemon
Verify the integration
- Open Memory System → Status and confirm dsh-mnemon, Mnemon Native, Runtime, Documents, and enabled Providers are listed
03DSH integration and capability boundaries
Native DSH plugin — a three-tier, pluggable, agent-driven memory control plane for DeepSeek Harness (Runtime Memory, Project Documents, and Memory Spaces backed by 9 long-term providers)
Three-tier memory model
Per-turn context (Runtime), complete narratives (Documents), cross-session facts (Memory Spaces)→Runtime: compact USER.md / MEMORY.md projection every turn; Documents: search-first with full Markdown on demand; Memory Spaces: bounded recall from active spaces
writes to local Mnemon store / external provider only when a Host-controlled task Agent accepts the writeNine pluggable long-term providers
A chosen Memory Space provider (Mnemon, OpenViking, Honcho, Mem0, Hindsight, Holographic, RetainDB, ByteRover, or Supermemory)→Cross-session facts, decisions, entities and relations retrieved on demand from the active space
may write to an external HTTP provider (e.g. Mem0, Honcho, Supermemory) depending on selectionSupervised write workflow
Remember / Save-to-memory actions and smart selection→A clean task Agent qualifies, routes, deduplicates, distills and writes behind Host controls; a routing receipt is saved
04Who is it for? When not to use it?
Good for
- DSH users and agents who need persistent, cross-session, multi-agent memory with pluggable long-term providers instead of a single flat store
Not for
- If a non-native provider is selected for Memory Spaces, writes/recall go to that external service (HTTP), whose deletion semantics and enumerable content vary by engine; dsh-mnemon never invents graph edges or deletion semantics for an engine that does not provide them.
05Compatibility, maintenance and safety notes
- The verified DSH version requires Node.js ^22.19.0 || >=24.0.0; Node 20 lacks host primitives the release depends on. The default Mnemon Native engine is a local CLI + SQLite (macOS / Linux via Go, Windows via release ZIP).
- Raw image bytes are never copied into memory; multimodal conversation messages keep their DSH attachment references through lifecycle processing, and Mnemon background jobs send text-only prompts.
- If a non-native provider is selected for Memory Spaces, writes/recall go to that external service (HTTP), whose deletion semantics and enumerable content vary by engine; dsh-mnemon never invents graph edges or deletion semantics for an engine that does not provide them.
MIT · actively maintained (latest release v0.2.16, 2026-08-22; verified against DSH 0.1.1-rc.2)
06Frequently asked questions
How do I install dsh-mnemon?
Install Mnemon Native first (`brew install --cask mnemon-dev/tap/mnemon` on macOS, or `go install github.com/mnemon-dev/mnemon@latest`), then install DSH 0.1.1-rc.2 (`npm install -g @deepseek-ai/dsh@0.1.1-rc.2`), and add the plugin with `dsh plugin --profile web add dsh-mnemon`.
Which memory providers can I use?
Nine: Mnemon (the official native local CLI + SQLite, prioritized), OpenViking, Honcho, Mem0, Hindsight, Holographic, RetainDB, ByteRover, and Supermemory. The third tier is replaceable; the first two tiers keep the same storage, workspace and interaction model regardless of provider.
What does a write actually do?
Remember / Save-to-memory runs a clean task Agent that qualifies, routes, deduplicates and distills behind Host controls, and only writes if accepted — saving a routing receipt. It never reuses or consumes the main conversation history.
Is my data safe with images and external providers?
Raw image bytes are never copied into memory; multimodal messages keep their DSH attachment references. For external providers, recall/writes go to that service and dsh-mnemon never invents graph edges or deletion semantics the engine doesn't provide.
07Related DSH workflows
memos
by memtensor
Self-evolving memory OS for LLM & AI Agents: ultra-persistent memory, hybrid-retrieval, and cross-task skill reuse, with 35.24% token savings and DeepSeek Harness support.
dsh-anchored-standard
by xiaobright
Two-phase DeepSeek Harness preset: Minimal-aligned bootstrap, then full Standard tools (Project2 98/99)
mem9
by mem9-ai
Unlimited memory for OpenClaw
mindmemos
by mindscale-noah
Gives dsh agents persistent cross-session memory: auto-recalls relevant user/task context before each turn and writes back lessons after each turn, evolving memory via schema learning and skill distillation
08Data and sources
The three-tier, pluggable, Agent-driven memory system for DeepSeek Harness.
dsh plugin --profile web add dsh-mnemon
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-08-23. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.