Skip to content

labring/sealos-skills

7019Last commit Aug 14, 2026

sealos-skills DSH plugin

Sealos Skills is a plugin-first skill pack centered on Sealos Cloud development and deployment. It helps an AI agent inspect a project, prepare deployment artifacts, connect Sealos Cloud databases and object storage, build or reuse container images, ship apps to Sealos Cloud, and view deployed resources in a local read-only canvas. It supports multiple AI agent tools including Codex, Claude Code, Qoder, and DeepSeek Harness.

How to install the sealos-skills DSH plugin

dsh plugin --profile web add github:labring/sealos-skills

Copying does not run this command. Review the repository and version before installing the sealos-skills DSH plugin.

sealos-skills DSH plugin data source

sealos-skills DSH plugin snapshot date: Aug 16, 2026

discovered

What the sealos-skills DSH plugin can do

  • Deploy projects to Sealos Cloud by inspecting project structure, building or reusing container images, and generating Sealos templates.
  • Create and manage Sealos Cloud databases (Postgres, etc.) and wire connection strings into local environment variables.
  • Create and manage S3-compatible object storage buckets on Sealos Cloud and wire credentials into local environment variables.
  • View a read-only canvas of deployed resources (requires verified deployment state).
  • Detect database signals (DATABASE_URL, Prisma, Drizzle, etc.) and object-storage signals (S3 env keys, AWS SDK usage, etc.) to automate setup.
  • Support multiple AI agent hosts: Codex, Claude Code, Qoder, DeepSeek Harness, and others through skills.sh or context-only extensions.

Where the sealos-skills DSH plugin fits

  • Deploy a GitHub repository to Sealos Cloud with a single command from your AI agent.
  • Create a cloud Postgres database for a local project and automatically wire the DATABASE_URL.
  • Create private S3 object storage for uploads and wire environment variables.
  • Inspect a deployed project's resources through a read-only canvas.
  • Migrate an existing Docker Compose project to Sealos Cloud using the docker-to-sealos skill.

Who the sealos-skills DSH plugin is for

  • Developers who want to deploy applications to Sealos Cloud via AI agents.
  • Teams using AI coding assistants (Codex, Claude Code, etc.) to streamline cloud deployment workflows.

sealos-skills DSH plugin limitations

  • Requires a Sealos Cloud account and access to a container registry (e.g., Docker Hub, GHCR) for actual deployments.
  • The default bash sandbox in DeepSeek Harness blocks writes outside the workspace; login writes to ~/.sealos/kubeconfig need sandbox_permissions: danger-full-access.
  • Only supported as a DeepSeek Harness profile bundle for the 'web' profile; no support for other profiles mentioned.
  • The canvas skill only works after a project has been deployed and verified state exists in .sealos/state.json.

sealos-skills DSH plugin: from the repository README

Quoted from the labring/sealos-skills README, the upstream source of the sealos-skills DSH plugin. Copyright remains with the original authors.

**English** | [简体中文](./readmes/README.zh-CN.md) | [繁體中文](./readmes/README.zh-TW.md) | [日本語](./readmes/README.ja.md) | [한국어](./readmes/README.ko.md) | [Español](./readmes/README.es.md) | [Français](./readmes/README.fr.md) | [Deutsch](./readmes/README.de.md) | [Português (Brasil)](./readmes/README.pt-BR.md) | [Русский](./readmes/README.ru.md) | [العربية](./readmes/README.ar.md) | [हिन्दी](./readmes/README.hi.md) | [Bahasa Indonesia](./readmes/README.id.md) <!-- README-I18N:END --> Deploy projects to [Sealos Cloud](https://sealos.io) from your AI agent. Sealos Skills is a plugin-first skill pack centered on Sealos Cloud development and deployment. It helps an AI agent inspect a project, prepare missing deployment artifacts, connect Sealos Cloud databases and object storage for development, build or reuse a container image, ship the app to Sealos Cloud, and view deployed resources in a local read-only canvas. The recommended Codex path is native Codex plugin installation. Cross-host plugin installs, `skills.sh`, and context-only extension hosts such as Gemini CLI and Qwen Code use the same root `skills/**` source. ## Quick Start ### Recommended: install in Codex Add this reposit

Read the full READMEThe repository declares no license. Check with the authors before using it.

sealos-skills DSH plugin questions

How do I install Sealos Skills in DeepSeek Harness?

First make sure `npx @deepseek-ai/dsh web` works. Then run `npx @deepseek-ai/dsh plugin --profile web add github:labring/sealos-skills`. After that, launch the web profile again with `npx @deepseek-ai/dsh web`. The eight root skills will appear in the session skill catalog.

What do I need to use Sealos Skills for deployment?

You need a Sealos Cloud account and access to a container registry (e.g., Docker Hub, GHCR). During the deployment flow, the agent will guide you through the Sealos login if needed. You don't need to set these up fully before starting the skill.

Why do I get a permission error when using Sealos Skills in DSH?

The default bash sandbox in DeepSeek Harness blocks writes outside the workspace. Sealos login writes to `~/.sealos/kubeconfig`, which is outside the workspace. You need to set `sandbox_permissions: danger-full-access` in your profile configuration to allow this.

Can I use Sealos Skills with other AI tools like Codex or Claude Code?

Yes, Sealos Skills is designed to work across multiple AI agent tools. For Codex, use `codex plugin marketplace add labring/sealos-skills` then `codex plugin add sealos@sealos`. For Claude Code, use `claude plugin marketplace add labring/sealos-skills` then `claude plugin install sealos@sealos`. The same skills are exposed as a managed package.

How do I view the resources deployed by Sealos Skills?

After a successful deployment, the agent creates a `.sealos/state.json` file with verified runtime evidence. You can then use the `sealos-canvas` skill (through the installed plugin entry point) to view a local read-only canvas of the deployed resources. This canvas is only available after deployment.