dsh-expert-mode
Agent & WorkflowAsher-2000/dsh-expert-mode
Expert-mode agent preset for DeepSeek Harness: Chief Coordinator plus 16 domain expert subagents for full-stack multi-agent team, auto-delegates tasks to best expert.
- agent-preset
- ai-agent
- deepseek-harness
- dsh-plugin
- expert-mode
- multi-agent
- subagent
README
๐ง DSH Expert Mode
โจ What it does
Install this preset and DSH automatically becomes a "Chief Coordinator" mode:
| Scenario | Behavior |
|---|---|
| Receives task | Identifies domain โ delegates to the best expert |
| Complex tasks | Dispatches multiple experts in parallel |
| Simple tasks | Coordinator handles directly โ no forced delegation |
| Task complete | Experts stay online for follow-up modifications |
No custom prompts to write. No multi-config to maintain. Just install and use.
๐ผ๏ธ Demo
๐งฉ 16 Experts
๐ฏ Full-Stack Core (6)
| Expert | Tool | Domain |
|---|---|---|
| ๐ฅ๏ธ Frontend Dev | expert_frontend_dev | Web frontend, React/Vue, CSS/UI |
| ๐ฅ๏ธ Backend Dev | expert_backend_dev | API, server logic, authentication |
| ๐๏ธ Database | expert_database | Schema design, SQL, optimization |
| ๐๏ธ Architect | expert_architect | System design, tech selection |
| ๐ ๏ธ DevOps | expert_devops | CI/CD, Docker, K8s, deployment |
| ๐งช QA Engineer | expert_qa_engineer | Testing strategy, automation |
๐ Security & Data (3)
| Expert | Tool | Domain |
|---|---|---|
| ๐ Security | expert_security | Code audit, vulnerabilities, hardening |
| ๐ Data Analyst | expert_data_analyst | Statistics, visualization, insights |
| ๐จ UI/UX Design | expert_uiux_design | Interface design, design systems |
๐ผ Business (7)
| Expert | Tool | Domain |
|---|---|---|
| ๐ Product Manager | expert_product_manager | PRD, requirements, competitor research |
| โ๏ธ Copywriter | expert_copywriter | Marketing copy, content creation |
| โ๏ธ Legal Review | expert_legal_review | Contract review, legal risk |
| ๐ฑ Social Media | expert_social_media | Multi-platform distribution |
| ๐ Growth Hacker | expert_growth | Growth strategy, A/B testing |
| ๐น Quant Finance | expert_quant_finance | Quantitative models, risk |
| ๐ฐ Finance | expert_finance | Financial analysis, budget |
๐ก๏ธ Features
| Feature | Description |
|---|---|
| ๐ฏ Smart Delegation | Auto-identifies task domain and routes to the best expert |
| ๐ Fast Track | Simple tasks handled directly โ no forced delegation |
| ๐ Five-Anchor Constraint | Prevents topic drift with per-turn self-check |
| ๐ค Cross Review | High-risk tasks get multi-expert independent review |
| ๐พ Experience Pool | Lessons learned are saved and injected next time |
| ๐ฌ Inter-Expert Bus | File-based message bus (bus.py): experts send/read directly, zero coordinator relay, P2P capable |
| โก Fault Recovery | Auto-retry on timeout, strategy switch on failure |
| ๐ Progressive Disclosure | Methodology injected on-demand, 28% token savings |
| ๐ Bilingual | Complete EN/ZH documentation |
๐ฆ Installation
Option A: dsh plugin add (won't crash, but โ does NOT activate expert mode)
# In DSH workspace
dsh plugin add github:Asher-2000/dsh-expert-mode
โ ๏ธ Important: this is an agent-preset plugin, not a Cordis service plugin.
dsh plugin addonly guarantees the bundle loads without crashing โ it does not mount expert mode into your sessions. The preset only works via Option B (~/.dsh/.agent-presets/). Use Option B. This entry is a compatibility placeholder only.
Option B: Manual install from GitHub (current)
Clone the repository, then copy the preset into DSH's agent-presets directory:
# 1. Clone anywhere
git clone https://github.com/Asher-2000/dsh-expert-mode.git
cd dsh-expert-mode
# 2. Copy the preset into DSH's agent-presets directory
mkdir -p ~/.dsh/.agent-presets/expert-mode
cp -r agent.cordis.yml preset.yml cordis.patch.yml ~/.dsh/.agent-presets/expert-mode/
# If you want the full methodology docs (methods/, experts/, comm/ bus), copy the whole tree:
# cp -r .expert-mode ~/.dsh/.agent-presets/expert-mode/
# 3. Restart DSH web, then select "ไธๅฎถๆจกๅผ" in the workspace preset selector
dsh web
Note:
~/.dsh/.agent-presets/is DSH's preset discovery directory. Each subdirectory = one preset. The preset name comes frompreset.yml'snamefield.
Then select "ไธๅฎถๆจกๅผ" in the workspace preset selector.
Optional: Cross-session memory (recommended)
The expert-mode preset itself does not register the cross-session memory service โ it is a HOST-PLANE plugin, and registering it inside a preset conflicts with the host composition (causing preset mount failure). To enable cross-session memory, install dsh-memory-connect separately into the host composition:
# 1. Clone the memory plugin
git clone https://github.com/Asher-2000/dsh-memory-connect.git
cd dsh-memory-connect
npm install github:Asher-2000/dsh-memory-connect#v0.4.0 # or place it into the dsh dependency tree manually
# 2. Register it in the host composition (e.g. append to ~/.dsh/profiles/web/cordis.patch.yml):
# - id: cross-session-memory
# name: '@deepseek-ai/dsh-memory-connect'
# config:
# path: ~/.dsh/memory.db
# openAt: startup
# 3. Restart DSH web
dsh web
โ ๏ธ Important: Do NOT add
@deepseek-ai/dsh-memory-connectinto this preset'sagent.cordis.yml. It is a HOST-PLANE plugin (injectssessions+systemPrompt); registering it inside the preset throwsservice has been registered at <cross-session-memory>, which makes the expert-mode preset fail to mount and the UI fall back to the default preset. This preset ships with an explanatory comment about it.
๐ Quick Start
- Install the plugin
- Select "ไธๅฎถๆจกๅผ" preset
- Ask any question โ the coordinator auto-delegates to the right expert
Example
User: ๅธฎๆ่ฎพ่ฎกไธไธช็จๆท่ฎค่ฏ็ณป็ป
Coordinator:
โ ่ฏๅซ้ขๅ: ๅ็ซฏๅผๅ + ๅฎๅ
จ
โ ๅงๆดพ Backend Dev: API ่ฎพ่ฎกใJWT ๅฎ็ฐ
โ ๅงๆดพ Security: ๅฎๅ
จๅฎก่ฎกใๆผๆด้ฒๆค
โ ๆฑๆป่พๅบๅฎๆดๆนๆก
๐๏ธ Architecture
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ Expert Mode Architecture โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ โ
โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ
โ โ Chief Coordinator (ๅ่ฐๅฎ) โ โ
โ โ โข Task analysis โข Domain identification โ โ
โ โ โข Expert routing โข Result aggregation โ โ
โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ
โ โ โ
โ โโโโโโโโโโโโโโโโโผโโโโโโโโโโโโโโโโ โ
โ โผ โผ โผ โ
โ โโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโ โ
โ โ Frontend โ โ Backend โ โ DevOps โ โ
โ โ Database โ โ Security โ โ QA โ โ
โ โ Architect โ โ ... โ โ ... โ โ
โ โโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโ โ
โ โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
๐ฌ Inter-Expert Communication Bus (v0.8.0)
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ File Message Bus (comm/bus.py) โ
โ .expert-mode/comm/mailboxes/<expert>/*.msg โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโค
โ โ
โ data-analyst โโsendโโโถ frontend-dev (direct, async) โ
โ copywriter โโsendโโโถ social-media (direct, async) โ
โ coordinator โโbroadcastโโโถ all experts (global sync) โ
โ expert A โโP2P subagentโโโถ expert B (synchronous) โ
โ โ
โ โข Zero relay: content flows between experts, NOT through โ
โ coordinator context โ
โ โข Durable: every message persisted as .msg file โ
โ โข Auditable: full log at comm/logs/bus.log โ
โ โข Commands: send / read / ack / broadcast / stats โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
Communication Modes:
| Mode | How | Use case |
|---|---|---|
| A. Relay | Expert A sends result โ Expert B reads | Sequential collaboration |
| B. Parallel | Experts send results to coordinator โ read --all | Independent collection |
| C. Broadcast | One message โ all mailboxes | Global state changes |
| D. Review | Experts send "agree/partial/disagree + reason" | Cross review |
| E. P2P | Expert spawns subagent for direct Q&A | Synchronous clarification |
๐ Documentation
| Document | Description |
|---|---|
| Communication Protocol | Inter-expert message bus protocol v1 |
| Expert Methods | 16 expert methodology docs |
| Experience Pool | Lessons learned per expert |
| README.zh.md | ไธญๆๆๆกฃ |
๐ค Contributing
- Fork the repository
- Create a feature branch
- Commit your changes
- Push to the branch
- Open a Pull Request
๐ License
MIT License - see LICENSE for details.
๐ Acknowledgments
- DeepSeek Harness - The core framework
- Cordis - Plugin system
- Awesome DSH Plugin - Community listing