DeepSeek Harness: Building and Extending AI Coding Agents Training Course

Introduction to DeepSeek Harness

• Definition of DeepSeek Harness and its ecosystem position
• The "everything is a plugin" design principle
• Overview of Web UI, CLI, and headless modes

Installation and Initial Setup

• Node.js prerequisites and package setup
• Starting the Web UI using npx
• Compilation and execution from source code

Model Setup

• Configuring the DeepSeek API key
• Model selection and switching procedures
• Integration with OpenAI-compatible endpoints

Workspaces and Sessions

• Workspace selection and management
• Session composer usage and conversation dynamics
• File reading, editing, and generation capabilities

Executing Coding Tasks

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

Permissions and Sandboxing

• Approval workflows and permission strategies
• Restricting file and command access
• Local and remote sandbox backend options

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 branching subagents
• Utilizing providers for other coding agents
• Distributing work across sessions

The Cordis Plugin Model

• Plugins as services, events, and effects
• Shared context and lifecycle management
• Reasons for the absence of a privileged core

Profiles, Bundles, and Settings

• Profile and bundle layering mechanisms
• Configuration adjustments via cordis.patch.yml
• Reviewing the initialized component tree

Core Subsystems

• The append-only session logging system
• Assembly of system prompts and tool schemas
• Agent loop mechanics and turn progression

Events and Extension Hooks

• Session, agent, and capability events
• Waterfall and interceptor event handling
• Aligning behaviors with extension points

Plugin Development

• Service registration via context keys
• Incorporating tools and model adapters
• Plugin publication and discovery processes

SDK, Headless Mode, and Automation

• Controlling the harness from external processes
• TypeScript and Python SDK integration
• CLI and headless automation techniques