Skip to content
Cursos/DeepSeek Harness (DSH)/DeepSeek Harness · Phase 2
Revisión pendiente desde 21 ago 20262 fuentes primarias

Installation and a safe first session

Your first DSH session should make the agent observable, not powerful. Use an empty workspace and a read-only task.

Qué vas a conseguir ahora

Una decisión o prueba aplicada a «Installation and a safe first session».

  1. 1Entiende el criterio
  2. 2Haz una práctica pequeña
  3. 3Guarda una evidencia
Evidencia

Una nota breve con qué hiciste, qué salió bien, qué falló y qué revisarías después.

Objetivos de aprendizaje
  • Prepare Node.js and a disposable workspace.
  • Start DSH's local interface.
  • Complete a read-only inspection with evidence.

Prepare a safe workspace

Use a fresh folder or a throwaway copy of a repository. Do not start with client data, production systems, secrets or uncommitted work.

The official README marks DSH as developer preview: it is moving quickly and can introduce breaking changes. Check the current documentation before using commands with a team.

Terminal
node -v

# npm path from the official README
npx @deepseek-ai/dsh web

# Usual URL printed by the command
http://127.0.0.1:3080

What happens when the Web UI starts

The dsh process uses the directory where you invoke it as the default filesystem location. Still, a fresh Web UI has no selected workspace until you add one.

This matters: the session composer remains unavailable until a workspace is selected. If you cannot type a task, the model is not broken; Choose workspace is probably missing.

  • Start DSH from the project folder or a test folder.
  • Open Choose workspace.
  • Add the directory where dsh was started.
  • Select that workspace before creating a session.

Run from source

Use the source path if you want to develop plugins, inspect Cordis or test patches. It is not required for basic use, but it matters for the advanced parts of this course.

Terminal
git clone https://github.com/deepseek-ai/deepseek-harness.git
cd deepseek-harness
pnpm install
pnpm run build
pnpm dsh web

The first task

Select the test workspace and choose the most restrictive permission level. The first mission is an inspection, not an edit.

The official guide starts with a repository summary task. Aulafy adds one extra constraint: require file citations so you can verify that the agent actually inspected the workspace.

  • Do not grant full-system access.
  • Ask for file references behind the summary.
  • Inspect the full Trajectory before approving actions.
  • Save the result and action trace as evidence.
Terminal
Summarize this workspace.
List main folders, configuration files and entry points.
Do not modify anything. Cite the files you used.

Official sources for this phase

Read the Web UI guide, README and model configuration guide before using DSH in a real environment.

  • Web UI quickstart: `https://deepseek-harness.github.io/deepseek-harness/en/guide/quickstart`
  • Official README: `https://github.com/deepseek-ai/deepseek-harness`
  • Model configuration guide: `https://deepseek-harness.github.io/deepseek-harness/en/guide/providers`
Comprueba que funciona. You have a test workspace, a read-only session and a saved summary with file references.
Guardar y reabrir el proyecto.
Visibility and limits come before automation.

Si has guardado la evidencia de esta lección, continúa con «DeepSeek models, providers and cost». Si no, repite la comprobación antes de avanzar.

Complete Aulafy mapSee how this lesson fits without leaving your path.

Complete Aulafy map

How all courses connect

This is not a checklist. Start with the foundation, choose an outcome, and go deeper only when your project needs more control.

  1. 1Understand
  2. 2Apply or build
  3. 3Operate with confidence
01

Choose an application

Turn the foundation into a visible outcome: a website, a business improvement, media, or an interactive experience.

Continue into the technical branch when you need to maintain code, data, or infrastructure.

02

Build with code

Prepare your environment, work with coding agents, and run models while keeping control of your projects.

This branch prepares you to design and operate reliable AI systems.

03

Take systems to production

Combine retrieval, agents, evaluation, security, deployment, and model adaptation when the problem requires it.

You do not need every course: choose the component your system needs and return as it grows.

View full catalogue