setup-assistant
About
The setup-assistant skill guides developers through installing, configuring, and performing first-time setup for the Claude Patent Creator MCP server. It provides a complete setup lifecycle from pre-installation checks to verification, including quick setup scripts and troubleshooting steps. Use it for initial installation, environment configuration, authentication setup, or when migrating to a new machine.
Documentation
Setup Assistant Skill
Expert system for installing, configuring, and setting up the Claude Patent Creator MCP server.
When to Use
Installing first time, setting up new environment, configuring authentication, troubleshooting installation, migrating to new machine, updating dependencies, verifying health.
Quick Setup (5 Minutes)
# 1. Create virtual environment
python -m venv venv
venv\Scripts\activate # Windows
source venv/bin/activate # Linux/macOS
# 2. Run installer
python install.py
# 3. Restart Claude Code
# 4. Test
# Ask Claude: "Search MPEP for claim definiteness"
Complete Setup Lifecycle
Phase 1: Pre-Installation Checks
Requirements:
- Python 3.9-3.12 (3.11 recommended)
- 8GB+ RAM (16GB recommended)
- 5GB free disk
- Optional: NVIDIA GPU with CUDA 12.x
Verify:
python --version # Should show 3.9-3.12
nvidia-smi # Optional: Check GPU
Phase 2: Virtual Environment
# Create venv
python -m venv venv
# Activate
venv\Scripts\activate # Windows
source venv/bin/activate # Linux/macOS
# Verify
which python # Should show venv path
Important: Always activate venv before running scripts!
Phase 3: Dependency Installation
Automated (Recommended):
python install.py
# Handles: PyTorch order, GPU detection, MCP registration
Manual (Advanced):
# Install PyTorch FIRST (critical!)
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu128
# Then install package (includes all dependencies)
pip install -e .
Why PyTorch order matters: If installed after sentence-transformers, you get CPU-only version.
Phase 4: BigQuery Authentication
# Authenticate
gcloud auth application-default login \
--scopes=https://www.googleapis.com/auth/cloud-platform
# Set project
gcloud config set project YOUR_PROJECT_ID
# Add to .env
echo "GOOGLE_CLOUD_PROJECT=YOUR_PROJECT_ID" >> .env
# Test
python scripts/test_bigquery.py
Phase 5: MCP Server Registration
Automated:
python install.py # Handles registration automatically
Manual:
# Get paths
patent-creator verify-config
# Register (use forward slashes!)
claude mcp add --transport stdio claude-patent-creator --scope user -- \
"C:/path/to/venv/Scripts/python.exe" \
"C:/path/to/mcp_server/server.py"
# Verify
claude mcp list
Critical: Restart Claude Code after registration!
Phase 6: Configuration Files
Create .env:
# Required
GOOGLE_CLOUD_PROJECT=your_project_id
ANTHROPIC_API_KEY=<YOUR_ANTHROPIC_API_KEY>
# Optional
PATENT_LOG_LEVEL=INFO
PATENT_LOG_FORMAT=human
PATENT_ENABLE_METRICS=true
# Windows only
CLAUDE_CODE_GIT_BASH_PATH=C:\dev\Git\bin\bash.exe
Phase 7: Health Check
patent-creator health
# Expected:
# [OK] Python version OK
# [OK] Dependencies installed
# [OK] PyTorch with CUDA
# [OK] MPEP index loaded
# [OK] BigQuery configured
# [OK] All systems operational
Phase 8: Testing
python scripts/test_install.py
python scripts/test_gpu.py
python scripts/test_bigquery.py
python scripts/test_analyzers.py
Phase 9: First Use Validation
Test each capability:
1. "Search MPEP for claim definiteness requirements"
2. "Search for patents about neural networks filed in 2024"
3. "Review these claims: 1. A system comprising..."
4. "Create a flowchart for user login process"
If all work -> Setup complete!
Common Setup Issues
| Issue | Solution |
|---|---|
| PyTorch CPU-only | Reinstall PyTorch FIRST |
| MCP not loading | Restart Claude Code, verify with claude mcp list |
| Path errors | Use forward slashes (/) not backslashes (\) |
| BigQuery fails | Re-auth: gcloud auth application-default login |
| Index not found | Build: patent-creator rebuild-index |
| Import errors | Activate venv |
Platform-Specific Notes
Windows
- PowerShell: Use
venv\Scripts\activate - Git Bash required for MCP commands
- Paths: Always forward slashes in MCP config
- CUDA: Install NVIDIA drivers + toolkit
Linux
- venv:
source venv/bin/activate - FAISS-GPU: Available on Linux only
- Permissions: May need sudo
macOS
- Apple Silicon: Use MPS (auto-detected)
- Intel: Use CPU or external GPU
- Homebrew: May need for dependencies
Update & Maintenance
Updating Dependencies
venv\Scripts\activate
pip install -e . --upgrade
python scripts/test_install.py
Rebuilding Index
patent-creator rebuild-index
# Wait 5-15 minutes
Re-registering MCP
claude mcp remove claude-patent-creator
python install.py
# Restart Claude Code
Quick Reference
Essential Commands
# Setup
python install.py
patent-creator health
claude mcp list
# Testing
python scripts/test_install.py
python scripts/test_gpu.py
python scripts/test_bigquery.py
# Maintenance
patent-creator rebuild-index
patent-creator verify-config
Critical Files
.env- Environment variablesrequirements.txt- Dependenciesmcp_server/index/- MPEP search indexpdfs/- MPEP PDF files
Quick Install
/plugin add https://github.com/RobThePCGuy/Claude-Patent-Creator/tree/main/setup-assistantCopy and paste this command in Claude Code to install this skill
GitHub 仓库
Related Skills
sglang
MetaSGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.
evaluating-llms-harness
TestingThis Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.
llamaguard
OtherLlamaGuard is Meta's 7-8B parameter model for moderating LLM inputs and outputs across six safety categories like violence and hate speech. It offers 94-95% accuracy and can be deployed using vLLM, Hugging Face, or Amazon SageMaker. Use this skill to easily integrate content filtering and safety guardrails into your AI applications.
langchain
MetaLangChain is a framework for building LLM applications using agents, chains, and RAG pipelines. It supports multiple LLM providers, offers 500+ integrations, and includes features like tool calling and memory management. Use it for rapid prototyping and deploying production systems like chatbots, autonomous agents, and question-answering services.
