How to run a first DeepSeek Harness task

The official Python SDK guide is one path from a clean checkout to a single finished task. It installs the published SDK, runs the shipped standalone minimal profile, and shows that same dsh profile called from your own program. Normal SDK execution needs no system Node.js. The installation includes a matching native runtime wheel and the dsh command.

Prerequisites

The guide requires all of the following:

• Python 3.10 or newer
• Git
• Linux x64, Linux arm64, macOS 14 or newer on arm64, or Windows x64
• A DeepSeek-compatible API endpoint and credential
• An isolated workspace and an isolated Harness home

Use a disposable checkout or container. The profile pins danger-full-access, so the platform-selected persistent shell can modify any path visible to the runtime. The example and the SDK never silently read ~/.dsh. You pass both locations as absolute paths.

Install the SDK

On Linux or macOS, clone the repository, create a virtual environment, and install the published package:

git clone https://github.com/deepseek-ai/deepseek-harness.git
cd deepseek-harness
python -m venv .venv
. .venv/bin/activate
python -m pip install deepseek-harness-sdk

On Windows PowerShell, the same install uses the Python launcher and a different activate script:

git clone https://github.com/deepseek-ai/deepseek-harness.git
Set-Location deepseek-harness
py -3.10 -m venv .venv
.venv\Scripts\Activate.ps1
python -m pip install deepseek-harness-sdk

Repository contributors who build the artifacts should use the Python contributor workflow named in the guide. A normal install stops at python -m pip install deepseek-harness-sdk.

Set the API key

Export the credential before the first run. When you need a compatible proxy endpoint, the guide also shows DEEPSEEK_BASE_URL, left commented in the example.

Linux or macOS:

export DEEPSEEK_API_KEY=sk-your-key-here
# export DEEPSEEK_BASE_URL=http://127.0.0.1:8000/v1

Windows PowerShell:

$env:DEEPSEEK_API_KEY = "sk-your-key-here"
# $env:DEEPSEEK_BASE_URL = "http://127.0.0.1:8000/v1"

sk-your-key-here is the guide’s placeholder. Put your credential in DEEPSEEK_API_KEY. Leave the base URL commented unless you need that proxy.

Run the checked-in example

From the repository root, the guide runs one task through the checked-in minimal example, with an explicit workspace, an explicit home, and a session id.

Linux or macOS:

python python/sdk/examples/minimal.py \
  --workspace /absolute/path/to/disposable-workspace \
  --dsh-home /absolute/path/to/example-dsh-home \
  --session-id example-001 \
  "Inspect the repository and fix the failing tests."

Windows PowerShell:

python python/sdk/examples/minimal.py `
  --workspace C:\work\disposable-workspace `
  --dsh-home C:\work\example-dsh-home `
  --session-id example-001 `
  "Inspect the repository and fix the failing tests."

Those paths are the guide’s placeholders. Substitute real absolute directories. The script prints the final assistant response. The selected home receives the generated sdk-minimal profile, installed plugins, and uncompressed JSONL session logs under sessions/.

In that example, the model comes from --model, then DSH_MODEL, then deepseek-v4-flash. The system prompt is DSH_SYSTEM_PROMPT, falling back to You are a helpful software engineer assistant.

Call DeepSeekHarness from your program

The same task can live in your own code. The guide constructs the client inside a context manager:

from pathlib import Path

from deepseek_harness import DeepSeekHarness

workspace = Path("/absolute/path/to/disposable-workspace").resolve()
dsh_home = Path("/absolute/path/to/example-dsh-home").resolve()
with DeepSeekHarness(
    provider="deepseek-official",
    model="deepseek-v4-flash",
    max_tokens=49_152,
    cwd=str(workspace),
    dsh_home=str(dsh_home),
    profile="sdk-minimal",
) as harness:
    result = harness.run(
        "Inspect the repository and fix the failing tests.",
        session_id="example-001",
    )

print(result.final_response)

That constructor sets provider deepseek-official, model deepseek-v4-flash, max_tokens of 49_152, profile sdk-minimal, the resolved workspace as cwd, and a separate dsh_home. harness.run takes the task string and session_id. print(result.final_response) is how the example surfaces the assistant text.

The SDK starts the bundled dsh --profile sdk-minimal process lazily and reuses it until context-manager exit. The profile, its persistent patch, the home patch, and any ordered patches tuple form the application configuration. There is no separate Python runtime bin or complete-config option.

Use a fresh home when profiles, plugins, credentials, settings, and sessions must be isolated. Use a fresh session id for independent work. Reuse a harness, home, and id only to continue the same durable conversation and session-owned resources.

What sdk-minimal includes

The guide’s table is the contract for this first task:

The profile’s sole bundle inserts the complete tree over an empty root and does not include dsh-base, so later base-profile tools cannot appear implicitly. It contains the SDK protocol, one environment-configured DeepSeek adapter, local execution, and persistence. Text file tools, settings, managed credentials, OTel telemetry, Web tools, subagents, and local instruction discovery are absent.

Automatic compaction preserves the complete session log. SDK prompt text does not execute /compact or other slash commands. Image reading requires an image-capable route.

The DeepSeek session-log contributor uploads unaccepted log events with DeepSeek requests by default, at most maxBytes (8 MiB) per request. Set session-log-deepseek.enabled: false in a profile patch to disable it.

Another profile is valid when it includes @deepseek-ai/dsh-sdk-app or another JSON-RPC server row. Missing server rows, unresolved plugins, and invalid patches fail during startup instead of falling back to another composition.

Plugins and the editor tool

You do not need plugins for the first run. The guide’s plugin commands are for dependencies and bundle layers that should persist in the home you already chose. Install pnpm only for this management command. Launching the installed SDK does not need it.

Linux or macOS:

export DSH_HOME=/absolute/path/to/example-dsh-home
dsh --profile sdk-minimal --dump-default-config >/dev/null
dsh plugin --profile sdk-minimal add file:/absolute/path/to/my-plugin-bundle

Windows PowerShell:

$env:DSH_HOME = "C:\work\example-dsh-home"
dsh --profile sdk-minimal --dump-default-config | Out-Null
dsh plugin --profile sdk-minimal add file:C:/work/my-plugin-bundle

The first command initializes the shipped standalone profile. The second forwards package management to pnpm, then records any installed package that exports a dsh.bundle layer. Edit $DSH_HOME/profiles/sdk-minimal/cordis.patch.yml for persistent row changes, or pass patch files from Python for per-launch changes.

The bundled runtime includes str_replace_editor, but sdk-minimal omits it from the default Cordis tree. To opt in, save this configuration as editor.patch.yml. The insert adds the editor using the existing filesystem provider:

- insert:
    - id: tool-str-replace-editor
      name: '@deepseek-ai/dsh-tool-str-replace-editor'

Pass patches=("/absolute/path/to/editor.patch.yml",) when constructing DeepSeekHarness(profile="sdk-minimal", ...), or put the patch in $DSH_HOME/profiles/sdk-minimal/cordis.patch.yml for persistent configuration. On the next runtime launch, model requests include str_replace_editor beside the default tools. The local filesystem provider uses the runtime working directory for relative paths. Like the minimal shell, it does not confine access to that directory. The same patch works with the standard sdk profile.

The installed wheel still packages the full web profile and frontend assets. Run dsh web against an explicit DSH_HOME when a Python SDK deployment also needs the browser application. web is a separate CLI application and cannot serve a Python SDK client.

What to run next

Create a disposable workspace and a separate Harness home, export DEEPSEEK_API_KEY, and run the checked-in example above with absolute paths and a session id you have not reused. After the script prints the final assistant response, move those same paths, the model string deepseek-v4-flash, and that session id into the DeepSeekHarness example if the task should live in your program. The guide points the exact tree to the bundle reference, the runnable program to the example reference, lifecycle and results to the Python SDK reference, and profile layering to the dsh CLI reference.

Sources

• Get started with the Python SDK