Senior Consult MCP
# Senior Consult MCP
> Let your AI agent consult top-tier models for expert technical guidance.
An MCP server that connects your AI agent to expert models (Claude, GPT, Gemini, DeepSeek, Z.ai) for architectural advice, code reviews, and complex problem-solving.
## Why?
Your AI agent is great, but sometimes it needs a second opinion. This MCP server lets it consult specialized models for high-stakes decisions without switching context.
## Tools
| Tool | Description |
| ------------------- | -------------------------------------- |
| ask_senior | General technical consultation |
| code_review | Security, performance & quality checks |
| architecture_advice | System design guidance |
| reset_history | Clear conversation memory |
## Quick Start
### 1. Configure MCP Client
Add to your MCP settings (Claude Desktop, Cursor, Windsurf, etc.):
```json
{
"mcpServers": {
"senior-consult": {
"command": "npx",
"args": ["-y", "senior-consult-mcp"],
"env": {
"ANTHROPIC_API_KEY": "your-key"
}
}
}
}
```
### 2. Use It
- "Ask senior to review this authentication flow for security issues"
- "Get architecture advice on microservices vs monolith for our scale"
- "Code review this function for potential memory leaks"
## Installation
**Option A: npx (Recommended)**
```bash
# No install needed - runs directly
npx -y senior-consult-mcp
```
**Option B: Global**
```bash
npm install -g senior-consult-mcp
```
**Option C: Local Development**
```bash
git clone https://github.com/silkyland/senior-consult-mcp.git
cd senior-consult-mcp
npm install
npm run build
npm start
```
## Environment Variables
### API Keys
Set at least one provider key:
```bash
ANTHROPIC_API_KEY=sk-ant-... # Claude
OPENAI_API_KEY=sk-... # GPT
GEMINI_API_KEY=... # Gemini
DEEPSEEK_API_KEY=sk-... # DeepSeek
ZAI_API_KEY=... # Z.ai (or ZHIPU_API_KEY)
OPENAI_COMPATIBLE_API_KEY=... # Custom endpoint
```
### Model Configuration (Optional)
Override default models:
```bash
# Claude
CLAUDE_MODEL=claude-sonnet-4-5
CLAUDE_URL=https://api.anthropic.com/v1/messages
# OpenAI
OPENAI_MODEL=gpt-4o
OPENAI_URL=https://api.openai.com/v1/chat/completions
# Gemini
GEMINI_MODEL=gemini-2.5-flash
# DeepSeek
DEEPSEEK_MODEL=deepseek-chat
DEEPSEEK_URL=https://api.deepseek.com/chat/completions
# Z.ai
ZAI_MODEL=glm-4-plus
ZAI_URL=https://open.bigmodel.cn/api/paas/v4/chat/completions
# Custom OpenAI-compatible
OPENAI_COMPATIBLE_MODEL=your-model
OPENAI_COMPATIBLE_URL=https://your-endpoint/v1/chat/completions
```
### Full Example Config
```json
{
"mcpServers": {
"senior-consult": {
"command": "npx",
"args": ["-y", "senior-consult-mcp"],
"env": {
"ANTHROPIC_API_KEY": "sk-ant-xxx",
"DEEPSEEK_API_KEY": "sk-xxx",
"GEMINI_API_KEY": "xxx",
"CLAUDE_MODEL": "claude-sonnet-4-5",
"DEEPSEEK_MODEL": "deepseek-reasoner"
}
}
}
}
```
## Features
- Multi-Provider - Claude, GPT, Gemini, DeepSeek, Z.ai, custom endpoints
- Smart Memory - Token-aware history with auto-trimming
- Optimized Prompts - Tuned for concise, actionable advice
- Zero Config - Works out of the box with sensible defaults
## Contributing
We welcome contributions! Here's how you can help:
### Ways to Contribute
- Report bugs - Open an issue with reproduction steps
- Suggest features - Share ideas in discussions
- Improve docs - Fix typos, add examples
- Submit PRs - Bug fixes, new providers, optimizations
### Development Setup
```bash
# Clone & install
git clone https://github.com/silkyland/senior-consult-mcp.git
cd senior-consult-mcp
npm install
# Build & test
npm run build
# Run locally
npm start
```
### Pull Request Process
1. Fork the repo
2. Create feature branch (git checkout -b feature/awesome)
3. Make changes & add tests
4. Commit (git commit -m 'Add awesome feature')
5. Push (git push origin feature/awesome)
6. Open PR with clear description
### Code Style
- TypeScript strict mode
- ESLint + Prettier
- Meaningful commit messages
- Tests for new features
## License
MIT
---
**Questions?** Open an issue or start a discussion.
**Like this project?** Give it a star to show support!
TDQS
Scored across 4 tools
The tools have overlapping purposes that could cause confusion, particularly between 'architecture_advice' and 'ask_senior' which both involve consulting a senior AI. However, 'code_review' and 'reset_history' are clearly distinct, and descriptions help differentiate the general consultation from specific architecture advice.
The naming follows a consistent snake_case pattern throughout, with clear verb_noun structures like 'get_architecture_advice' implied. Minor deviations exist, such as 'ask_senior' being less descriptive than others, but overall the conventions are readable and predictable.
With 4 tools, the count is reasonable for a consultation server, covering core functions like advice, review, and history management. It's slightly lean but well-scoped, with each tool earning its place without feeling excessive or insufficient for the domain.
The tool set covers key consultation aspects like advice and code review, but there are notable gaps. For example, there's no tool for follow-up queries or iterative feedback, which could limit agent workflows. The surface is functional but not fully comprehensive for a senior consultation domain.