A native DeepSeek Harness memory plugin (dsh plugin) that gives agents persistent, cross-session recall via system-prompt injection and three memory tools.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-09-03
License
AGPL-3.0
01What can it help you accomplish?
Give DeepSeek Harness (dsh) agents long-term memory so they recall users, sessions, and projects across turns
Peer representations, session context, hybrid search results, and natural-language insights retrieved from Honcho and injected into the model prompt
Developers building stateful agents on DeepSeek Harness who want persistent, cross-session memory
Query and inspect what an agent knows about a person or project without retraining the model
Prompt-ready context blocks (session.context().to_openai / .to_anthropic) plus low-latency static representations
Teams shipping personalized, multi-session AI experiences on DeepSeek Harness
02How to install into DeepSeek Harness
Prerequisites
- A Honcho API key from app.honcho.dev (or a self-hosted Honcho server)
- Run `honcho init` once to write the key to `~/.honcho/config.json` for every integration
Installation steps
- 01
Install the native plugin into DeepSeek Harness: `dsh plugin --profile <name> add @honcho-ai/dsh-honcho`
$ dsh plugin --profile <name> add @honcho-ai/dsh-honcho
- 02
Open a session in DeepSeek Harness; Honcho injects memory into the system prompt and captures new session events
Verify the integration
- Run `/honcho` inside DeepSeek Harness to check the plugin’s status
03DSH integration and capability boundaries
Native Cordis plugin installed via the `dsh plugin` CLI — injects memory into the system prompt and captures session events
Memory injection into the system prompt
session event feed from DeepSeek Harness→relevant peer/session memory injected into the system prompt
captures new information from the session event feed and stores it in HonchoSession event capture
messages and events from the active session→stored interaction history in Honcho
sends session messages to Honcho (managed api.honcho.dev or your self-hosted server)Three memory tools (honcho_search, honcho_chat, honcho_remember)
natural-language queries from the model→search results, chat answers, and remembered facts
Prompt-ready context
a session or peer→context blocks via session.context(...).to_openai(...) / .to_anthropic(...)
04Who is it for? When not to use it?
Good for
- Developers building stateful agents on DeepSeek Harness who want persistent, cross-session memory
- Teams shipping personalized, multi-session AI experiences on DeepSeek Harness
Not for
- As a memory plugin it captures new information from the session event feed and sends it to a Honcho server — by default the managed api.honcho.dev endpoint, unless you self-host and point the SDK at your own base_url. Review where your session data lands before enabling it.
05Compatibility, maintenance and safety notes
- As a memory plugin it captures new information from the session event feed and sends it to a Honcho server — by default the managed api.honcho.dev endpoint, unless you self-host and point the SDK at your own base_url. Review where your session data lands before enabling it.
- Honcho is open source under AGPL-3.0; self-hosting and the CLI are covered by that license, so check copyleft obligations before embedding it in a closed-source product.
AGPL-3.0 · DeepSeek Harness plugin published by Plastic Labs
06Frequently asked questions
How do I install Honcho into DeepSeek Harness?
Run `dsh plugin --profile <name> add @honcho-ai/dsh-honcho` — it is a native Cordis plugin. After that, Honcho injects memory into the system prompt and captures new information from the session event feed.
Do I need an API key or account?
Yes. Get a key at app.honcho.dev (or run your own server), then run `honcho init` once to write it to `~/.honcho/config.json`; that one key configures every Honcho integration.
How do I confirm the plugin is working?
Run `/honcho` inside DeepSeek Harness to check the plugin’s status.
Does my session data leave my machine?
By default it goes to the managed api.honcho.dev endpoint. You can self-host the FastAPI server (`honcho start --setup` or Docker) and point the SDK at `http://localhost:8000` to keep data local.
What tools does the model get?
Three tools: honcho_search (hybrid BM25 + vector search), honcho_chat (ask what Honcho knows about a peer), and honcho_remember (store new facts).
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.
mirage
by strukto-ai
The World's First Virtual Terminal for AI Agents
reme
by agentscope-ai
ReMe: Memory Management Kit for Agents - Remember Me, Refine Me.
memsearch
by zilliztech
A persistent, unified memory layer for all your AI agents (e.g. Claude Code, Codex, DSH), backed by Markdown and Milvus.
08Data and sources
A native Cordis plugin. It injects memory into the system prompt and captures new information from the session event fee…
dsh plugin --profile <name> add @honcho-ai/dsh-honcho
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-09-03. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.