Capture conversations into L0–L3 memory and inject relevant items before DSH model steps. Background distillation uses LLM calls; default global scope exposes memory across workspaces.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-10-24
License
MIT
01What can it help you accomplish?
Carry relevant memory into later DSH model steps
Layered conversation records, scene summaries and persona memory injected before model steps
DSH users who want conversation memory and automatic recall
02How to install into DeepSeek Harness
Prerequisites
- Node.js 22.16 or later.
Installation steps
- 01
$ dsh plugin --profile web add dsh-prime-memory
- 02
$ dsh --profile web --dump-config
- 03
Restart the DSH Web Host manually after installation.
Verify the integration
- Confirm dsh-prime-memory appears in --dump-config; after restarting, check ~/.dsh/memory/ for conversations/, records/, scenes/ and memory.db.
Rollback
- Run dsh plugin --profile web remove dsh-prime-memory and restart. Data remains in ~/.dsh/memory/.
03DSH integration and capability boundaries
Native DSH event capture, distillation and recall injection
Layered L0–L3 capture and recall injection
Per-session memory-mode controls
04Who is it for? When not to use it?
Good for
- DSH users who want conversation memory and automatic recall
Not for
- Background distillation makes LLM calls and records token costs by provider/model in SQLite.
- The default global scope is visible across workspaces. Workspace scope isolates the work family.
- The README marks DSH 0.1.5-rc.2 untested and requiring regression checks; do not treat every listed release as verified.
05Compatibility, maintenance and safety notes
- Background distillation makes LLM calls and records token costs by provider/model in SQLite.
- The default global scope is visible across workspaces. Workspace scope isolates the work family.
- The README marks DSH 0.1.5-rc.2 untested and requiring regression checks; do not treat every listed release as verified.
- Vector retrieval is off by default; the default uses full-text search.
README author claims; not independently audited.
06Frequently asked questions
How can I check installation in DSH?
Use --dump-config to confirm the plugin, then restart the Web Host and check the documented memory directories and memory.db.
Does DSH memory stay inside one workspace?
Not by default. Global scope is shared across workspaces; workspace scope isolates the work family.
Does it make additional DSH model calls?
Yes. Distillation calls the host LLM and records token costs. Vector retrieval is separately disabled by default.
Does removing the DSH plugin delete memory?
No. The documented uninstall command leaves ~/.dsh/memory/ in place.
07Related DSH workflows
weknora
by tencent
Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.
honcho
by plastic-labs
Memory library for building stateful agents
webnovel-writer
by lingfengqaq
A long-form Chinese web-novel writing assistant for Claude Code and DeepSeek Harness, designed to reduce memory loss and hallucinations in serialized projects of up to 2 million characters.
mirage
by strukto-ai
The World's First Virtual Terminal for AI Agents
08Data and sources
**DeepSeek Harness 的分层蒸馏记忆插件：对话在后台自动完成 L0 捕获 → L1 原子记忆 → L2 场景整合 → L3 画像蒸馏，模型每一步前自动把相关记忆注入上下文。**
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-10-24. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.