> For the complete documentation index, see [llms.txt](https://evo-5.gitbook.io/evo/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://evo-5.gitbook.io/evo/host-integrations/deepseek-harness.md).

# DeepSeek Harness

evo integrates with DeepSeek Harness as a Cordis plugin, providing automatic recall and reflection plus a web-based Settings → Memory panel.

## Quick Install

For an npx-managed Harness profile:

```bash
./install_evo_dsps.sh
```

This builds the package, installs a local link through `dsh plugin`, and activates evo as a DSH bundle. The default profile is `web`; override with `DSH_PROFILE=tui` or another profile name.

It prefers a `dsh` already on your `PATH`, so the CLI that writes the profile is the one that later boots it. Set `DSH_PACKAGE` to pin a released version through `npx` instead.

### Re-running repairs the profile

A profile can end up carrying evo under several names at once — an alias left by an older install command, or a former published name such as `evo-memory`. Each resolves to the same `cordis.patch.yml`, so DSH applies that patch once per name, inserting every evo plugin id more than once, and the profile stops booting with an error that names neither the alias nor the duplication.

Re-running the installer removes them:

```
evo: removing stale evo installs from profile 'web': evo evo-memory
```

A name is only removed when it is provably evo: it resolves to a package whose manifest name is `@tiz36/evo`, or it is one of evo's former names and resolves to nothing. A name that resolves to some *other* package is reported and left alone — `evo` is a real name on the registry.

The bundle uses `deepseek-official` and `deepseek-v4-flash` by default; override at launch with `EVO_PROVIDER` and `EVO_MODEL`.

## Manual Configuration

Add the plugins to your `cordis.patch.yml` after the Harness LLM and system-prompt services:

```yaml
- id: evo
  name: evo/cordis
  config: {}

- id: evo-deepseek
  name: evo/deepseek
  config:
    provider: deepseek-official
    model: deepseek-chat
    recallLimit: 40
    maxContextChars: 6000
    reflect: true

# Optional: web panel carrier
- id: evo-web
  name: evo
  config: {}
```

A copyable fragment is available in [`examples/cordis.yml`](https://github.com/TIZ36/evo/blob/main/examples/cordis.yml).

## How It Works

The adapter uses official Harness extension points:

| Extension Point          | Role                                                             |
| ------------------------ | ---------------------------------------------------------------- |
| `system-prompt/assemble` | Recalls global, project, and session memory into dynamic context |
| `session/event`          | Observes successful `turn/end` events and reflects activity      |
| `ctx.llm.stream()`       | Runs reflection and consolidation through the configured model   |

Interrupted, aborted, rejected, and failed turns are never reflected.

## Web Panel

The package ships a web client that adds a **Memory** page to Harness GUI Settings:

* Memory list with kind tabs and search
* Recent reflect/consolidate activity log
* Actions to consolidate a scope or force workspace re-import

The composer tool row also gets an **evo** chip that pulses during reflection. Clicking it opens Settings → Memory.

The client half is carried by the `evo-web` row in `cordis.patch.yml`. Restart Harness after upgrading so the new boot graph includes the panel.

## Configuration Options

| Option            | Default | Description                                    |
| ----------------- | ------- | ---------------------------------------------- |
| `provider`        | —       | Harness provider for reflection                |
| `model`           | —       | Model for reflection and consolidation         |
| `recallLimit`     | `40`    | Maximum memories to consider per recall        |
| `maxContextChars` | `6000`  | Character budget for injected context          |
| `reflect`         | `true`  | Set `false` to recall only, never write        |
| `workspaceImport` | `true`  | Set `false` to skip automatic workspace import |


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://evo-5.gitbook.io/evo/host-integrations/deepseek-harness.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.