Code
How deepseek-ai/deepseek-harness is laid out, and where to start for common changes.
There is no import map for this report yet.
Where to start reading
- Repository root
- Repository root
- Likely public API
- Likely public API
- Likely start-up file
- Likely start-up file
- Suggested starting pointAI
- Suggested starting pointAI
Find files for a task
AIAdd a new plugin/skill
- Create a new package under packages/[feature-name]/ following the monorepo structure
- Implement plugin class extending core plugin interfaces; expose via packages/core abstractions
- Register plugin capabilities in MCP protocol if exposing tools to LLM
- Add plugin to host initialization configuration in packages/host
Add LLM provider support
- Implement provider adapter in packages/llm following abstraction interfaces
- Integrate with credential storage for API keys and auth tokens
Add a new tool/environment capability
- Choose capability type: browser (packages/browser-use), computer (packages/computer-use), terminal (packages/terminal), or custom
- Implement environment plugin and expose actions as MCP tools
- If requiring isolation, configure sandbox constraints in packages/sandbox
- Register tool availability in host configuration
Add a Web UI component or page
- Create component in apps/web following React/Vue patterns (check existing structure)
- Connect to session/context via packages/session and packages/context hooks
- Communicate with host via API layer in packages/api if needed
Change the public API
Trace runtime behavior
Add or update tests
Change build or automation
RepoPilot lists up to 500 files, so a file missing here may still exist.
- Files marked “likely” are guessed from their names.