DeepSeek Harness: Building and Extending AI Coding Agents Training Course

Introduction to DeepSeek Harness

• Defining DeepSeek Harness and its application scope
• The design philosophy of "everything is a plugin"
• Overview of Web UI, CLI, and headless modes

Installation and Initial Execution

• Node.js prerequisites and package setup
• Starting the Web UI via npx
• Building and running the project from source code

Model Configuration

• Configuring the DeepSeek API key
• Choosing and switching between models
• Integrating OpenAI-compatible endpoints

Workspaces and Session Management

• Selecting and navigating workspaces
• Utilizing the session composer and conversation flow
• File reading, editing, and generation tasks

Executing Coding Tasks

• Repository summarization and code navigation
• Command execution and task delegation
• Maintaining an active task plan

Permissions and Sandbox Environment

• Approval workflows and permission policies
• Restricting file and command access
• Configuring local and remote sandbox backends

Native Tools and Integrations

• Filesystem, terminal, and shell utilities
• Support for the Model Context Protocol
• Integration with the Language Server Protocol

Subagents and Task Delegation

• Creating and forking subagents
• Using product providers for other coding agents
• Distributing work across a session

The Cordis Plugin Model

• Plugins functioning as services, events, and effects
• Shared context and lifecycle management
• Elimination of a privileged core structure

Profiles, Bundles, and Configuration

• Profile and bundle layering strategies
• Configuration patching via cordis.patch.yml
• Reviewing the boot-up tree structure

Core Subsystems

• The append-only session logging system
• Assembly of system prompts and tool schemas
• The agent loop and turn-by-turn flow

Events and Extension Points

• Session, agent, and capability events
• Waterfall and interceptor event handling
• Mapping specific behaviors to extension points

Plugin Development

• Registering services on context keys
• Implementing new tools and model adapters
• Plugin publication and discovery mechanisms

SDK, Headless Mode, and Automation

• Controlling the harness from external processes
• Utilizing TypeScript and Python SDKs
• Automating workflows via CLI and headless mode