DeepSeek Harness: Building and Extending AI Coding Agents Training Course

Introduction to DeepSeek Harness

• Overview of DeepSeek Harness and its role in the ecosystem
• The "everything is a plugin" architectural principle
• Operational modes: Web UI, CLI, and headless

Setup and Initial Execution

• Node.js prerequisites and installing dependencies
• Starting the Web interface using npx
• Compiling and executing from source code

Model Setup

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

Workspaces and Session Management

• Selecting and managing active workspaces
• Utilizing the session composer for interactive flows
• File operations: reading, editing, and generating

Executing Coding Tasks

• Repository summarization and code navigation
• Command execution and task delegation
• Maintaining structured task plans

Permissions and Sandbox Security

• Approval workflows and permission policies
• Restricting file and command access
• Local and remote sandbox implementations

Native Tools and Integrations

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

Subagents and Task Delegation

• Creating and forking subagents
• Providers for alternative coding agents
• Distributing work across sessions

The Cordis Plugin Architecture

• Plugins defined as services, events, and effects
• Shared context and lifecycle management
• Rationale behind the absence of a privileged core

Profiles, Bundles, and Configuration Management

• Profile and bundle layering strategies
• Modifying settings via cordis.patch.yml
• Analyzing the initialized component tree

Core System Components

• Append-only session logging mechanisms
• Assembly of system prompts and tool schemas
• Agent loop logic and turn processing

Event Handling and Extension Points

• Session, agent, and capability triggers
• Waterfall and interceptor event patterns
• Aligning behaviors with specific extension points

Plugin Development

• Registering services using context keys
• Implementing custom tools and model adapters
• Releasing and locating plugins

SDKs, Headless Operations, and Automation

• Controlling the harness from external processes
• Using TypeScript and Python SDKs
• Automating via CLI and headless modes