dsh-project-brain
Run the following command in DeepSeek Harness:
dsh plugin install yj-liuzepeng/dsh-project-brain
Paste the following prompt into your AI chat to install this plugin:
Install the plugin in DeepSeek Harness by running dsh plugin install yj-liuzepeng/dsh-project-brain; the source repository is at https://github.com/yj-liuzepeng/dsh-project-brain
About this plugin
AI coding sessions are short-lived, but real software projects are not. Architectural decisions, production bug fixes, in-progress work, and critical path files get scattered across conversations, forcing every new session to rediscover the repository from scratch. dsh-project-brain turns that context into a local, workspace-scoped knowledge layer that grows more useful as the project evolves.
Core capabilities include detecting languages, frameworks, entry points, and dependency structure to generate a conceptual architecture report covering project purpose, architectural style, component responsibilities, runtime flows, and risks; writing decisions, requirements, bugs, lessons, TODOs, and timeline events through 13 project_* tools; and restoring high-value memories, active TODOs, and recent activity into new sessions via the DSH system prompt. The default BM25 retriever requires zero configuration, while optional hybrid keyword-and-vector retrieval provides a semantic upgrade. All memory write and architecture analysis paths degrade gracefully to fully local operation when the AI is unavailable.
It is designed for developers who work daily in DSH, want project context to persist across sessions, and prefer to avoid external memory services or extra model keys. Workspaces are strictly isolated so one project memory is never guessed from another, and all data lives in a local .project-brain/ directory that survives plugin upgrades and is never deleted on uninstall.
Screenshots
Use Cases
- New sessions automatically restore high-value decisions, active TODOs, and recent project activity
- Quickly detect the tech stack of an unfamiliar repo and generate a conceptual architecture report with reading order
- Persist decision rationale, bug-fix insights, and lessons across sessions to avoid repeating mistakes
Best For
- Engineers who develop daily in DSH and want project context to persist across sessions
- Teams that need strict per-workspace memory isolation without external memory services
- Developers who want local architecture analysis and BM25 retrieval as a fallback when AI is unavailable
Related Plugins
Memory layer for coding agents that indexes local session history and auto-recalls relevant context before edits, commands, and failures, with no manual search needed.
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.