| Name | instrmcp JSON |
| Version |
2.0.1
JSON |
| download |
| home_page | None |
| Summary | MCP server suite for physics laboratory instrumentation control |
| upload_time | 2025-10-21 00:57:32 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.10 |
| license | MIT License
Copyright (c) 2024-2025 Jiaqi Cai <jiaqic@mit.edu>
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
|
| keywords |
mcp
model-context-protocol
qcodes
instrumentation
physics
laboratory
measurement
automation
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# InstrMCP: Instrumentation Control MCP Server
[](https://pypi.org/project/instrmcp/)
[](https://www.python.org/downloads/)
[](https://opensource.org/licenses/MIT)
[](https://github.com/anthropics/mcp)
[](https://instrmcp.readthedocs.io/en/latest/?badge=latest)
[](https://github.com/caidish/instrMCP/actions/workflows/tests.yml)
[](https://github.com/caidish/instrMCP/actions/workflows/lint.yml)
MCP server suite for quantum device physics laboratory's instrumentation control, enabling Large Language Models to interact directly with physics instruments and measurement systems through QCodes and JupyterLab.
https://github.com/user-attachments/assets/1d4d6e42-138c-4f49-90ef-803eb6c01488
## Features
- **Full QCodes Integration**: Built-in support for all QCodes instrument drivers
- **Database Integration**: Read-only access to QCodes databases with intelligent code generation
- **MeasureIt Templates**: Comprehensive measurement pattern library and code generation
- **JupyterLab Native**: Seamless integration with JupyterLab
- **Dynamic Tool Creation**: Create custom MCP tools at runtime using LLM-powered tool registration
- **Safe mode**: Read-only mode with optional unsafe execution
- **CLI**: Easy server management with `instrmcp` command
- **MCP**: Standard Model Context Protocol for LLM integration
- The MCP has been tested to work with Claude Desktop, Claude Code, and Codex CLI
## Quick Start
### Installation
**From PyPI (Recommended):**
```bash
pip install instrmcp
```
**That's it!** QCodes, JupyterLab, and all dependencies are automatically installed. The JupyterLab extension is automatically enabled (no Node.js or rebuild required).
**From Source (For Development):**
```bash
git clone https://github.com/caidish/instrMCP.git
cd instrMCP
pip install -e .
# Run setup to enable JupyterLab extension (only needed for editable install)
instrmcp-setup
# Set required environment variable for development
# For macOS/Linux:
export instrMCP_PATH="$(pwd)"
echo 'export instrMCP_PATH="'$(pwd)'"' >> ~/.zshrc # or ~/.bashrc
source ~/.zshrc
# For Windows (PowerShell):
$env:instrMCP_PATH = (Get-Location).Path
[System.Environment]::SetEnvironmentVariable('instrMCP_PATH', (Get-Location).Path, 'User')
```
### Extension: MeasureIt
MeasureIt provides comprehensive measurement pattern templates for common physics experiments.
**Installation:**
```bash
pip install qmeasure
```
**Important Notes:**
- Import as `measureit` (not `qmeasure`): `import measureit`
- Python 3.8+ required
- For source installation or advanced configuration, see the [MeasureIt GitHub repository](https://github.com/nanophys/MeasureIt)
**Enable in InstrMCP:**
```python
# In Jupyter notebook
%mcp_option add measureit
%mcp_restart
```
### Usage
#### Loading InstrMCP in Jupyter
```bash
# Start JupyterLab
jupyter lab
```
In a Jupyter notebook cell:
```python
# Load InstrMCP extension
%load_ext instrmcp.extensions
# Start MCP server
%mcp_start
# Check status
%mcp_status
# Enable unsafe mode (code execution)
%mcp_unsafe
# Enable optional features (restart required)
%mcp_option add measureit database
%mcp_restart
```
#### CLI Server Management
```bash
# Start standalone servers
instrmcp jupyter --port 3000 # Jupyter MCP server
instrmcp jupyter --port 3000 --unsafe # With unsafe mode
instrmcp qcodes --port 3001 # QCodes station server
# Configuration and info
instrmcp config # Show configuration paths
instrmcp version # Show version
instrmcp --help # Show all commands
```
## Documentation
- **[Architecture](docs/ARCHITECTURE.md)** - Technical architecture, package structure, MCP tools and resources
- **[Troubleshooting](docs/TROUBLESHOOTING.md)** - Common issues and solutions
- **[Development Guide](docs/DEVELOPMENT.md)** - Development setup, testing, code quality, contributing
## Configuration Example
Station configuration uses standard YAML format:
```yaml
# instrmcp/config/data/default_station.yaml
instruments:
mock_dac:
driver: qcodes.instrument_drivers.mock.MockDAC
name: mock_dac_1
enable: true
```
Configuration is automatic! The system auto-detects installation paths. For custom setups:
```bash
# View current configuration
instrmcp config
# Custom config file (optional)
mkdir -p ~/.instrmcp
echo "custom_setting: value" > ~/.instrmcp/config.yaml
```
## Claude Desktop Integration
InstrMCP provides seamless integration with Claude Desktop for AI-assisted laboratory instrumentation control.
### Quick Setup (2 Steps)
1. **Run Automated Setup**:
```bash
cd /path/to/your/instrMCP
./claudedesktopsetting/setup_claude.sh
```
2. **Restart Claude Desktop** completely and test with: *"What MCP tools are available?"*
**Manual Setup Alternative:**
```bash
# 1. Copy and edit configuration
cp claudedesktopsetting/claude_desktop_config.json ~/Library/Application\ Support/Claude/claude_desktop_config.json
# 2. Edit the copied file - replace placeholders with actual paths:
# /path/to/your/python3 → $(which python3)
# /path/to/your/instrMCP → $(pwd)
```
See [`claudedesktopsetting/README.md`](claudedesktopsetting/README.md) for detailed setup instructions and troubleshooting.
## Claude Code Integration
Claude Code supports local MCP servers via STDIO. Use the provided launcher to connect:
```bash
# Add instrMCP as MCP Server
claude mcp add instrMCP --env instrMCP_PATH=$instrMCP_PATH \
--env PYTHONPATH=$instrMCP_PATH \
-- $instrMCP_PATH/venv/bin/python \
$instrMCP_PATH/claudedesktopsetting/claude_launcher.py
# Verify connection
/mcp
```
**Prerequisites:**
- Ensure `instrMCP_PATH` environment variable is set
- Have a Jupyter server running with the instrMCP extension loaded
- MCP server started in Jupyter with `%mcp_start`
## Codex CLI Integration
Codex expects MCP servers over STDIO. Use the Codex launcher to proxy STDIO calls to your HTTP MCP server.
**Configuration:**
- command: `python`
- args: `["/path/to/your/instrMCP/codexsetting/codex_launcher.py"]`
- env:
- `JUPYTER_MCP_HOST=127.0.0.1`
- `JUPYTER_MCP_PORT=8123`
## V2.0.0 Features (Current Release)
### 1. Resource Discovery Tool
The `mcp_list_resources()` tool helps LLMs discover and effectively use MCP resources:
**Features:**
- **Comprehensive Resource Listing**: All available MCP resources with URIs, descriptions, and use cases
- **Context-Aware**: Only shows resources based on enabled options (core, MeasureIt, database)
- **Resources vs Tools Guidance**: Educates LLMs on when to use read-only resources vs active tools
- **Common Patterns**: Examples like "Check available_instruments → Use qcodes_instrument_info"
- **First-Use Recommendation**: Use this tool FIRST to discover what context is available
**Example Response:**
```json
{
"total_resources": 8,
"resources": [
{
"uri": "resource://available_instruments",
"name": "Available Instruments",
"use_when": "Need to know what instruments exist BEFORE calling qcodes_instrument_info",
"example": "Check this first to see instrument names..."
}
],
"guidance": {
"resources_vs_tools": {
"resources": "Provide READ-ONLY reference data, templates, and documentation",
"tools": "Perform ACTIVE operations like reading live values, executing code"
},
"when_to_use_resources": [
"Before using tools - check available_instruments first",
"For code templates - get MeasureIt examples",
"For configuration - check database_config"
]
}
}
```
### 2. Visual Diff Consent for Notebook Patching
Enhanced `notebook_apply_patch` with visual diff dialog showing exactly what will change:
**Features:**
- **Visual Diff Display**: Red deletions, green additions, context lines
- **Pattern Warning**: Prominent alert if old_text not found in cell
- **Change Statistics**: Shows chars removed/added and delta
- **Consent Workflow**: "Decline" | "Allow" | "Always Allow" buttons
- **Battle-Tested Diffing**: Uses industry-standard `diff` library (v8.0.2) from GitHub/GitLab
**Example:** When LLM calls `notebook_apply_patch(old_text="x = 10", new_text="x = 20")`, user sees:
```diff
- x = 10 ← Red background with strikethrough
+ x = 20 ← Green background
```
### 3. Line Range Parameters for Context Management
Control LLM context window consumption with line range selection:
**Features:**
- **line_start** / **line_end** parameters (default: lines 1-100)
- **Automatic Bounds Clamping**: Invalid ranges safely handled
- **Truncation Metadata**: Returns `total_lines`, `truncated` flag
- **Context Window Savings**: Prevents large cells from consuming excessive tokens
**Example:**
```python
# Get only first 50 lines of a large cell
get_editing_cell(line_start=1, line_end=50)
# Get lines 100-200 for focused analysis
get_editing_cell(line_start=100, line_end=200)
```
### 4. Dynamic Tool Creation
Create custom MCP tools at runtime using LLM-powered tool registration:
```python
# In Jupyter with instrMCP loaded in unsafe mode
# LLM can create tools dynamically using meta-tools:
dynamic_register_tool(
name="analyze_data",
source_code="def analyze_data(x): return x * 2",
capabilities=["cap:numpy", "cap:custom.analysis"], # Freeform labels
parameters=[{"name": "x", "type": "number", "description": "Input", "required": true}]
)
```
**Features:**
- **6 Meta-Tools**: `register`, `update`, `revoke`, `list`, `inspect`, `registry_stats`
- **Consent UI**: User approval required for tool registration/updates (shows full source code)
- **Freeform Capability Labels**: Tag tools with descriptive capabilities for discovery
- **Persistent Registry**: Tools saved to `~/.instrmcp/registry/` and reloaded on server start
- **Audit Trail**: All tool operations logged to `~/.instrmcp/audit/tool_audit.log`
- **Auto JSON Correction**: Optional LLM-powered JSON error fixing (opt-in via `%mcp_option auto_correct_json`)
**Capability Labels** (v2.0.0):
Capabilities are freeform documentation labels - NOT enforced security boundaries. Use any descriptive string:
- Suggested format: `cap:library.action` (e.g., `cap:numpy.array`, `cap:qcodes.read`)
- Used for discovery, filtering, and transparency in consent UI
- No validation - flexibility for LLMs to describe tool dependencies
- Future: Enforcement layer planned for v3.0.0
See [Dynamic Tools Quickstart](docs/DYNAMIC_TOOLS_QUICKSTART.md) for details.
### Testing & Quality
- **464 tests** (463 passed, 1 skipped)
- **Zero linter errors** across all modules
- **Code formatted** with black
- **CI/CD passing** on all workflows
## V3.0.0 Roadmap
- **Capability Enforcement**: Security boundaries based on capability taxonomy
- Support RedPitaya
- Support Raspberry Pi for outdated instruments
- Integrating lab wiki knowledge base for safety rails
- More LLM integration examples
## License
MIT License - see [LICENSE](LICENSE) file.
## Contributing
We welcome contributions! See our [Development Guide](docs/DEVELOPMENT.md) for details on:
- Setting up development environment
- Running tests
- Code quality standards
- Contribution guidelines
## Links
- [Documentation](https://instrmcp.readthedocs.io)
- [Issues](https://github.com/caidish/instrMCP/issues)
- [QCodes](https://qcodes.github.io/Qcodes/)
- [Model Context Protocol](https://github.com/anthropics/mcp)
Raw data
{
"_id": null,
"home_page": null,
"name": "instrmcp",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": "Jiaqi Cai <jiaqic@mit.edu>",
"keywords": "mcp, model-context-protocol, qcodes, instrumentation, physics, laboratory, measurement, automation",
"author": null,
"author_email": "Jiaqi Cai <jiaqic@mit.edu>",
"download_url": "https://files.pythonhosted.org/packages/29/4e/c772fede04c7c056177955e5ce27a2c4a3d5e0e862f5d40d3b1f42a395a1/instrmcp-2.0.1.tar.gz",
"platform": null,
"description": "# InstrMCP: Instrumentation Control MCP Server\n\n[](https://pypi.org/project/instrmcp/)\n[](https://www.python.org/downloads/)\n[](https://opensource.org/licenses/MIT)\n[](https://github.com/anthropics/mcp)\n[](https://instrmcp.readthedocs.io/en/latest/?badge=latest)\n[](https://github.com/caidish/instrMCP/actions/workflows/tests.yml)\n[](https://github.com/caidish/instrMCP/actions/workflows/lint.yml)\n\nMCP server suite for quantum device physics laboratory's instrumentation control, enabling Large Language Models to interact directly with physics instruments and measurement systems through QCodes and JupyterLab.\n\nhttps://github.com/user-attachments/assets/1d4d6e42-138c-4f49-90ef-803eb6c01488\n\n## Features\n\n- **Full QCodes Integration**: Built-in support for all QCodes instrument drivers\n- **Database Integration**: Read-only access to QCodes databases with intelligent code generation\n- **MeasureIt Templates**: Comprehensive measurement pattern library and code generation\n- **JupyterLab Native**: Seamless integration with JupyterLab\n- **Dynamic Tool Creation**: Create custom MCP tools at runtime using LLM-powered tool registration\n- **Safe mode**: Read-only mode with optional unsafe execution\n- **CLI**: Easy server management with `instrmcp` command\n- **MCP**: Standard Model Context Protocol for LLM integration\n- The MCP has been tested to work with Claude Desktop, Claude Code, and Codex CLI\n## Quick Start\n\n### Installation\n\n**From PyPI (Recommended):**\n\n```bash\npip install instrmcp\n```\n\n**That's it!** QCodes, JupyterLab, and all dependencies are automatically installed. The JupyterLab extension is automatically enabled (no Node.js or rebuild required).\n\n**From Source (For Development):**\n\n```bash\ngit clone https://github.com/caidish/instrMCP.git\ncd instrMCP\npip install -e .\n\n# Run setup to enable JupyterLab extension (only needed for editable install)\ninstrmcp-setup\n\n# Set required environment variable for development\n# For macOS/Linux:\nexport instrMCP_PATH=\"$(pwd)\"\necho 'export instrMCP_PATH=\"'$(pwd)'\"' >> ~/.zshrc # or ~/.bashrc\nsource ~/.zshrc\n\n# For Windows (PowerShell):\n$env:instrMCP_PATH = (Get-Location).Path\n[System.Environment]::SetEnvironmentVariable('instrMCP_PATH', (Get-Location).Path, 'User')\n```\n\n### Extension: MeasureIt\n\nMeasureIt provides comprehensive measurement pattern templates for common physics experiments.\n\n**Installation:**\n```bash\npip install qmeasure\n```\n\n**Important Notes:**\n- Import as `measureit` (not `qmeasure`): `import measureit`\n- Python 3.8+ required\n- For source installation or advanced configuration, see the [MeasureIt GitHub repository](https://github.com/nanophys/MeasureIt)\n\n**Enable in InstrMCP:**\n```python\n# In Jupyter notebook\n%mcp_option add measureit\n%mcp_restart\n```\n\n### Usage\n\n#### Loading InstrMCP in Jupyter\n\n```bash\n# Start JupyterLab\njupyter lab\n```\n\nIn a Jupyter notebook cell:\n\n```python\n# Load InstrMCP extension\n%load_ext instrmcp.extensions\n\n# Start MCP server\n%mcp_start\n\n# Check status\n%mcp_status\n\n# Enable unsafe mode (code execution)\n%mcp_unsafe\n\n# Enable optional features (restart required)\n%mcp_option add measureit database\n%mcp_restart\n```\n\n#### CLI Server Management\n\n```bash\n# Start standalone servers\ninstrmcp jupyter --port 3000 # Jupyter MCP server\ninstrmcp jupyter --port 3000 --unsafe # With unsafe mode\ninstrmcp qcodes --port 3001 # QCodes station server\n\n# Configuration and info\ninstrmcp config # Show configuration paths\ninstrmcp version # Show version\ninstrmcp --help # Show all commands\n```\n\n## Documentation\n\n- **[Architecture](docs/ARCHITECTURE.md)** - Technical architecture, package structure, MCP tools and resources\n- **[Troubleshooting](docs/TROUBLESHOOTING.md)** - Common issues and solutions\n- **[Development Guide](docs/DEVELOPMENT.md)** - Development setup, testing, code quality, contributing\n\n## Configuration Example\n\nStation configuration uses standard YAML format:\n\n```yaml\n# instrmcp/config/data/default_station.yaml\ninstruments:\n mock_dac:\n driver: qcodes.instrument_drivers.mock.MockDAC\n name: mock_dac_1\n enable: true\n```\n\nConfiguration is automatic! The system auto-detects installation paths. For custom setups:\n\n```bash\n# View current configuration\ninstrmcp config\n\n# Custom config file (optional)\nmkdir -p ~/.instrmcp\necho \"custom_setting: value\" > ~/.instrmcp/config.yaml\n```\n\n## Claude Desktop Integration\n\nInstrMCP provides seamless integration with Claude Desktop for AI-assisted laboratory instrumentation control.\n\n### Quick Setup (2 Steps)\n\n1. **Run Automated Setup**:\n```bash\ncd /path/to/your/instrMCP\n./claudedesktopsetting/setup_claude.sh\n```\n\n2. **Restart Claude Desktop** completely and test with: *\"What MCP tools are available?\"*\n\n**Manual Setup Alternative:**\n```bash\n# 1. Copy and edit configuration\ncp claudedesktopsetting/claude_desktop_config.json ~/Library/Application\\ Support/Claude/claude_desktop_config.json\n\n# 2. Edit the copied file - replace placeholders with actual paths:\n# /path/to/your/python3 \u2192 $(which python3)\n# /path/to/your/instrMCP \u2192 $(pwd)\n```\n\nSee [`claudedesktopsetting/README.md`](claudedesktopsetting/README.md) for detailed setup instructions and troubleshooting.\n\n## Claude Code Integration\n\nClaude Code supports local MCP servers via STDIO. Use the provided launcher to connect:\n\n```bash\n# Add instrMCP as MCP Server\nclaude mcp add instrMCP --env instrMCP_PATH=$instrMCP_PATH \\\n --env PYTHONPATH=$instrMCP_PATH \\\n -- $instrMCP_PATH/venv/bin/python \\\n $instrMCP_PATH/claudedesktopsetting/claude_launcher.py\n\n# Verify connection\n/mcp\n```\n\n**Prerequisites:**\n- Ensure `instrMCP_PATH` environment variable is set\n- Have a Jupyter server running with the instrMCP extension loaded\n- MCP server started in Jupyter with `%mcp_start`\n\n## Codex CLI Integration\n\nCodex expects MCP servers over STDIO. Use the Codex launcher to proxy STDIO calls to your HTTP MCP server.\n\n**Configuration:**\n- command: `python`\n- args: `[\"/path/to/your/instrMCP/codexsetting/codex_launcher.py\"]`\n- env:\n - `JUPYTER_MCP_HOST=127.0.0.1`\n - `JUPYTER_MCP_PORT=8123`\n\n## V2.0.0 Features (Current Release)\n\n### 1. Resource Discovery Tool\nThe `mcp_list_resources()` tool helps LLMs discover and effectively use MCP resources:\n\n**Features:**\n- **Comprehensive Resource Listing**: All available MCP resources with URIs, descriptions, and use cases\n- **Context-Aware**: Only shows resources based on enabled options (core, MeasureIt, database)\n- **Resources vs Tools Guidance**: Educates LLMs on when to use read-only resources vs active tools\n- **Common Patterns**: Examples like \"Check available_instruments \u2192 Use qcodes_instrument_info\"\n- **First-Use Recommendation**: Use this tool FIRST to discover what context is available\n\n**Example Response:**\n```json\n{\n \"total_resources\": 8,\n \"resources\": [\n {\n \"uri\": \"resource://available_instruments\",\n \"name\": \"Available Instruments\",\n \"use_when\": \"Need to know what instruments exist BEFORE calling qcodes_instrument_info\",\n \"example\": \"Check this first to see instrument names...\"\n }\n ],\n \"guidance\": {\n \"resources_vs_tools\": {\n \"resources\": \"Provide READ-ONLY reference data, templates, and documentation\",\n \"tools\": \"Perform ACTIVE operations like reading live values, executing code\"\n },\n \"when_to_use_resources\": [\n \"Before using tools - check available_instruments first\",\n \"For code templates - get MeasureIt examples\",\n \"For configuration - check database_config\"\n ]\n }\n}\n```\n\n### 2. Visual Diff Consent for Notebook Patching\nEnhanced `notebook_apply_patch` with visual diff dialog showing exactly what will change:\n\n**Features:**\n- **Visual Diff Display**: Red deletions, green additions, context lines\n- **Pattern Warning**: Prominent alert if old_text not found in cell\n- **Change Statistics**: Shows chars removed/added and delta\n- **Consent Workflow**: \"Decline\" | \"Allow\" | \"Always Allow\" buttons\n- **Battle-Tested Diffing**: Uses industry-standard `diff` library (v8.0.2) from GitHub/GitLab\n\n**Example:** When LLM calls `notebook_apply_patch(old_text=\"x = 10\", new_text=\"x = 20\")`, user sees:\n```diff\n- x = 10 \u2190 Red background with strikethrough\n+ x = 20 \u2190 Green background\n```\n\n### 3. Line Range Parameters for Context Management\nControl LLM context window consumption with line range selection:\n\n**Features:**\n- **line_start** / **line_end** parameters (default: lines 1-100)\n- **Automatic Bounds Clamping**: Invalid ranges safely handled\n- **Truncation Metadata**: Returns `total_lines`, `truncated` flag\n- **Context Window Savings**: Prevents large cells from consuming excessive tokens\n\n**Example:**\n```python\n# Get only first 50 lines of a large cell\nget_editing_cell(line_start=1, line_end=50)\n\n# Get lines 100-200 for focused analysis\nget_editing_cell(line_start=100, line_end=200)\n```\n\n### 4. Dynamic Tool Creation\nCreate custom MCP tools at runtime using LLM-powered tool registration:\n\n```python\n# In Jupyter with instrMCP loaded in unsafe mode\n# LLM can create tools dynamically using meta-tools:\ndynamic_register_tool(\n name=\"analyze_data\",\n source_code=\"def analyze_data(x): return x * 2\",\n capabilities=[\"cap:numpy\", \"cap:custom.analysis\"], # Freeform labels\n parameters=[{\"name\": \"x\", \"type\": \"number\", \"description\": \"Input\", \"required\": true}]\n)\n```\n\n**Features:**\n- **6 Meta-Tools**: `register`, `update`, `revoke`, `list`, `inspect`, `registry_stats`\n- **Consent UI**: User approval required for tool registration/updates (shows full source code)\n- **Freeform Capability Labels**: Tag tools with descriptive capabilities for discovery\n- **Persistent Registry**: Tools saved to `~/.instrmcp/registry/` and reloaded on server start\n- **Audit Trail**: All tool operations logged to `~/.instrmcp/audit/tool_audit.log`\n- **Auto JSON Correction**: Optional LLM-powered JSON error fixing (opt-in via `%mcp_option auto_correct_json`)\n\n**Capability Labels** (v2.0.0):\nCapabilities are freeform documentation labels - NOT enforced security boundaries. Use any descriptive string:\n- Suggested format: `cap:library.action` (e.g., `cap:numpy.array`, `cap:qcodes.read`)\n- Used for discovery, filtering, and transparency in consent UI\n- No validation - flexibility for LLMs to describe tool dependencies\n- Future: Enforcement layer planned for v3.0.0\n\nSee [Dynamic Tools Quickstart](docs/DYNAMIC_TOOLS_QUICKSTART.md) for details.\n\n### Testing & Quality\n- **464 tests** (463 passed, 1 skipped)\n- **Zero linter errors** across all modules\n- **Code formatted** with black\n- **CI/CD passing** on all workflows\n\n## V3.0.0 Roadmap\n\n- **Capability Enforcement**: Security boundaries based on capability taxonomy\n- Support RedPitaya\n- Support Raspberry Pi for outdated instruments\n- Integrating lab wiki knowledge base for safety rails\n- More LLM integration examples\n\n## License\n\nMIT License - see [LICENSE](LICENSE) file.\n\n## Contributing\n\nWe welcome contributions! See our [Development Guide](docs/DEVELOPMENT.md) for details on:\n- Setting up development environment\n- Running tests\n- Code quality standards\n- Contribution guidelines\n\n## Links\n\n- [Documentation](https://instrmcp.readthedocs.io)\n- [Issues](https://github.com/caidish/instrMCP/issues)\n- [QCodes](https://qcodes.github.io/Qcodes/)\n- [Model Context Protocol](https://github.com/anthropics/mcp)\n",
"bugtrack_url": null,
"license": "MIT License\n \n Copyright (c) 2024-2025 Jiaqi Cai <jiaqic@mit.edu>\n \n Permission is hereby granted, free of charge, to any person obtaining a copy\n of this software and associated documentation files (the \"Software\"), to deal\n in the Software without restriction, including without limitation the rights\n to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n copies of the Software, and to permit persons to whom the Software is\n furnished to do so, subject to the following conditions:\n \n The above copyright notice and this permission notice shall be included in all\n copies or substantial portions of the Software.\n \n THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\n SOFTWARE.\n ",
"summary": "MCP server suite for physics laboratory instrumentation control",
"version": "2.0.1",
"project_urls": {
"Changelog": "https://github.com/instrmcp/instrMCP/blob/main/CHANGELOG.md",
"Documentation": "https://instrmcp.readthedocs.io",
"Homepage": "https://github.com/instrmcp/instrMCP",
"Issues": "https://github.com/instrmcp/instrMCP/issues",
"Repository": "https://github.com/instrmcp/instrMCP.git"
},
"split_keywords": [
"mcp",
" model-context-protocol",
" qcodes",
" instrumentation",
" physics",
" laboratory",
" measurement",
" automation"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "ceb352d08ac1bd27314e434e0ab4175110027471dbcd3bf3ac9a3a471bce8814",
"md5": "b4572cbbb8a0f3aab3da492367bc5a46",
"sha256": "bfc1e08b43a0437183b0c0ac122e14c0ff8324761bb97b0b7cb33b57934e0caf"
},
"downloads": -1,
"filename": "instrmcp-2.0.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "b4572cbbb8a0f3aab3da492367bc5a46",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10",
"size": 158051,
"upload_time": "2025-10-21T00:57:26",
"upload_time_iso_8601": "2025-10-21T00:57:26.677666Z",
"url": "https://files.pythonhosted.org/packages/ce/b3/52d08ac1bd27314e434e0ab4175110027471dbcd3bf3ac9a3a471bce8814/instrmcp-2.0.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "294ec772fede04c7c056177955e5ce27a2c4a3d5e0e862f5d40d3b1f42a395a1",
"md5": "17471256e2a1cad0418de23d7af90cba",
"sha256": "d5919aa4bd85c338bd44a3f9207e79ba98ec33ec9e5ca3497516d8c560e81dda"
},
"downloads": -1,
"filename": "instrmcp-2.0.1.tar.gz",
"has_sig": false,
"md5_digest": "17471256e2a1cad0418de23d7af90cba",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 13995528,
"upload_time": "2025-10-21T00:57:32",
"upload_time_iso_8601": "2025-10-21T00:57:32.770348Z",
"url": "https://files.pythonhosted.org/packages/29/4e/c772fede04c7c056177955e5ce27a2c4a3d5e0e862f5d40d3b1f42a395a1/instrmcp-2.0.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-10-21 00:57:32",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "instrmcp",
"github_project": "instrMCP",
"github_not_found": true,
"lcname": "instrmcp"
}
