hanzo-mcp 0.1.25__tar.gz → 0.1.30__tar.gz

{hanzo_mcp-0.1.25 → hanzo_mcp-0.1.30}/LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2025 Hanzo Industries Inc
+Copyright (c) 2025 hanzoai
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal
@@ -18,4 +18,4 @@ 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.
+SOFTWARE.

{hanzo_mcp-0.1.25 → hanzo_mcp-0.1.30}/PKG-INFO

@@ -1,9 +1,13 @@
 Metadata-Version: 2.4
 Name: hanzo-mcp
-Version: 0.1.25
-Summary: MCP server for accessing Hanzo APIs and Platform capabilities
-Author-email: Hanzo <dev@hanzo.ai>
+Version: 0.1.30
+Summary: MCP implementation of Hanzo capabilities
+Author-email: Hanzo Industries Inc <dev@hanzo.ai>
 License: MIT
+Project-URL: Homepage, https://github.com/hanzoai/mcp
+Project-URL: Bug Tracker, https://github.com/hanzoai/mcp/issues
+Project-URL: Documentation, https://github.com/hanzoai/mcp/blob/main/README.md
+Keywords: mcp,claude,hanzo,code,agent
 Classifier: Programming Language :: Python :: 3
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Operating System :: OS Independent
@@ -13,8 +17,9 @@ License-File: LICENSE
 Requires-Dist: mcp>=1.3.0
 Requires-Dist: httpx>=0.27.0
 Requires-Dist: uvicorn>=0.23.1
-Requires-Dist: mcp
-Requires-Dist: pytest
+Requires-Dist: openai>=1.50.0
+Requires-Dist: python-dotenv>=1.0.0
+Requires-Dist: litellm>=1.40.14
 Provides-Extra: dev
 Requires-Dist: pytest>=7.0.0; extra == "dev"
 Requires-Dist: pytest-cov>=4.1.0; extra == "dev"
@@ -29,15 +34,18 @@ Requires-Dist: twisted; extra == "test"
 Provides-Extra: performance
 Requires-Dist: ujson>=5.7.0; extra == "performance"
 Requires-Dist: orjson>=3.9.0; extra == "performance"
+Provides-Extra: publish
+Requires-Dist: twine>=4.0.2; extra == "publish"
+Requires-Dist: build>=1.0.3; extra == "publish"
 Dynamic: license-file
 
 # Hanzo MCP
 
-Hanzo AI + Platform capabilities via the Model Context Protocol (MCP).
+An implementation of Hanzo capabilities using the Model Context Protocol (MCP).
 
 ## Overview
 
-This project provides an MCP server that enables access to Hanzo APIs and Platform capabilities, as well as providing development tools for managing and improving projects. By leveraging the Model Context Protocol, this server enables seamless integration with various MCP clients including Claude Desktop, allowing LLMs to directly access Hanzo's platform functionality.
+This project provides an MCP server that implements Hanzo-like functionality, allowing Claude to directly execute instructions for modifying and improving project files. By leveraging the Model Context Protocol, this implementation enables seamless integration with various MCP clients including Claude Desktop.
 
 ![example](./doc/example.gif)
 
@@ -49,6 +57,8 @@ This project provides an MCP server that enables access to Hanzo APIs and Platfo
 - **File Operations**: Manage files with proper security controls through shell commands
 - **Code Discovery**: Find relevant files and code patterns across your project
 - **Project Analysis**: Understand project structure, dependencies, and frameworks
+- **Agent Delegation**: Delegate complex tasks to specialized sub-agents that can work concurrently
+- **Multiple LLM Provider Support**: Configure any LiteLLM-compatible model for agent operations
 - **Jupyter Notebook Support**: Read and edit Jupyter notebooks with full cell and output handling
 
 ## Tools Implemented
@@ -65,84 +75,27 @@ This project provides an MCP server that enables access to Hanzo APIs and Platfo
 | `run_command`          | Execute shell commands (also used for directory creation, file moving, and directory listing) |
 | `run_script`           | Execute scripts with specified interpreters                                                   |
 | `script_tool`          | Execute scripts in specific programming languages                                             |
-| `project_analyze_tool` | Analyze project structure and dependencies                                                   |
-| `read_notebook`        | Extract and read source code from all cells in a Jupyter notebook with outputs               |
-| `edit_notebook`        | Edit, insert, or delete cells in a Jupyter notebook                                         |
+| `project_analyze_tool` | Analyze project structure and dependencies                                                    |
+| `read_notebook`        | Extract and read source code from all cells in a Jupyter notebook with outputs                |
+| `edit_notebook`        | Edit, insert, or delete cells in a Jupyter notebook                                           |
 | `think`                | Structured space for complex reasoning and analysis without making changes                    |
+| `dispatch_agent`       | Launch one or more agents that can perform tasks using read-only tools concurrently           |
 
 ## Getting Started
 
-### Usage
-
-#### Configuring Claude Desktop
-
-You can run it with `uvx run hanzo-mcp` without installation. Configure Claude Desktop to use this server by adding the following to your Claude Desktop configuration file:
-
-```json
-{
-  "mcpServers": {
-    "hanzo": {
-      "command": "uvx",
-      "args": [
-        "--from",
-        "hanzo-mcp",
-        "hanzo-mcp",
-        "--allow-path",
-        "/path/to/your/project"
-      ]
-    }
-  }
-}
-```
-
-Make sure to replace `/path/to/your/project` with the actual path to the project you want Claude to access.
-
-#### Advanced Configuration Options
-
-You can customize the server using other options:
-
-```json
-{
-  "mcpServers": {
-    "hanzo": {
-      "command": "uvx",
-      "args": [
-        "--from",
-        "hanzo-mcp",
-        "hanzo-mcp",
-        "--allow-path",
-        "/path/to/project",
-        "--name",
-        "custom-hanzo",
-        "--transport",
-        "stdio"
-      ]
-    }
-  }
-}
-```
-
-### Configuring Claude Desktop System Prompt
-
-To get the best experience with Hanzo MCP, you need to add the provided system prompt to your Claude Desktop client. This system prompt guides Claude through a structured workflow for interacting with Hanzo platform services and managing project files.
-
-Follow these steps:
+### Quick Install
 
-1. Locate the system prompt file in this repository at `doc/system_prompt`
-2. Open your Claude Desktop client
-3. Create a new project or open an existing one
-4. Navigate to the "Project instructions" section in the Claude Desktop sidebar
-5. Copy the contents of `doc/system_prompt` and paste it into the "Project instructions" section
-6. Replace `{project_path}` with the actual absolute path to your project
+```bash
+# Install using uv
+uv pip install hanzo-mcp
 
-The system prompt provides Claude with:
+# Or using pip
+pip install hanzo-mcp
+```
 
-- A structured workflow for analyzing and modifying code
-- Best practices for project exploration and analysis
-- Guidelines for development, refactoring, and quality assurance
-- Special formatting instructions for mathematical content
+For detailed installation and configuration instructions, please refer to [INSTALL.md](./doc/INSTALL.md).
 
-This step is crucial as it enables Claude to follow a consistent approach when helping you with code modifications.
+Of course, you can also read [USEFUL_PROMPTS](./doc/USEFUL_PROMPTS.md) for some inspiration on how to use hanzo-mcp.
 
 ## Security
 
@@ -155,6 +108,48 @@ This implementation follows best practices for securing access to your filesyste
 
 ## Development
 
+### Setup Development Environment
+
+```bash
+# Clone the repository
+git clone https://github.com/hanzoai/mcp.git
+cd mcp
+
+# Install Python 3.13 using uv
+make install-python
+
+# Setup virtual environment and install dependencies
+make setup
+
+# Or install with development dependencies
+make install-dev
+```
+
+### Testing
+
+```bash
+# Run tests
+make test
+
+# Run tests with coverage
+make test-cov
+```
+
+### Building and Publishing
+
+```bash
+# Build package
+make build
+
+# Publish to PyPI
+make publish
+
+# Publish to Test PyPI
+make publish-test
+```
+
+### Contributing
+
 To contribute to this project:
 
 1. Fork the repository

{hanzo_mcp-0.1.25 → hanzo_mcp-0.1.30}/README.md

@@ -1,10 +1,10 @@
 # Hanzo MCP
 
-Hanzo AI + Platform capabilities via the Model Context Protocol (MCP).
+An implementation of Hanzo capabilities using the Model Context Protocol (MCP).
 
 ## Overview
 
-This project provides an MCP server that enables access to Hanzo APIs and Platform capabilities, as well as providing development tools for managing and improving projects. By leveraging the Model Context Protocol, this server enables seamless integration with various MCP clients including Claude Desktop, allowing LLMs to directly access Hanzo's platform functionality.
+This project provides an MCP server that implements Hanzo-like functionality, allowing Claude to directly execute instructions for modifying and improving project files. By leveraging the Model Context Protocol, this implementation enables seamless integration with various MCP clients including Claude Desktop.
 
 ![example](./doc/example.gif)
 
@@ -16,6 +16,8 @@ This project provides an MCP server that enables access to Hanzo APIs and Platfo
 - **File Operations**: Manage files with proper security controls through shell commands
 - **Code Discovery**: Find relevant files and code patterns across your project
 - **Project Analysis**: Understand project structure, dependencies, and frameworks
+- **Agent Delegation**: Delegate complex tasks to specialized sub-agents that can work concurrently
+- **Multiple LLM Provider Support**: Configure any LiteLLM-compatible model for agent operations
 - **Jupyter Notebook Support**: Read and edit Jupyter notebooks with full cell and output handling
 
 ## Tools Implemented
@@ -32,84 +34,27 @@ This project provides an MCP server that enables access to Hanzo APIs and Platfo
 | `run_command`          | Execute shell commands (also used for directory creation, file moving, and directory listing) |
 | `run_script`           | Execute scripts with specified interpreters                                                   |
 | `script_tool`          | Execute scripts in specific programming languages                                             |
-| `project_analyze_tool` | Analyze project structure and dependencies                                                   |
-| `read_notebook`        | Extract and read source code from all cells in a Jupyter notebook with outputs               |
-| `edit_notebook`        | Edit, insert, or delete cells in a Jupyter notebook                                         |
+| `project_analyze_tool` | Analyze project structure and dependencies                                                    |
+| `read_notebook`        | Extract and read source code from all cells in a Jupyter notebook with outputs                |
+| `edit_notebook`        | Edit, insert, or delete cells in a Jupyter notebook                                           |
 | `think`                | Structured space for complex reasoning and analysis without making changes                    |
+| `dispatch_agent`       | Launch one or more agents that can perform tasks using read-only tools concurrently           |
 
 ## Getting Started
 
-### Usage
-
-#### Configuring Claude Desktop
-
-You can run it with `uvx run hanzo-mcp` without installation. Configure Claude Desktop to use this server by adding the following to your Claude Desktop configuration file:
-
-```json
-{
-  "mcpServers": {
-    "hanzo": {
-      "command": "uvx",
-      "args": [
-        "--from",
-        "hanzo-mcp",
-        "hanzo-mcp",
-        "--allow-path",
-        "/path/to/your/project"
-      ]
-    }
-  }
-}
-```
-
-Make sure to replace `/path/to/your/project` with the actual path to the project you want Claude to access.
-
-#### Advanced Configuration Options
-
-You can customize the server using other options:
-
-```json
-{
-  "mcpServers": {
-    "hanzo": {
-      "command": "uvx",
-      "args": [
-        "--from",
-        "hanzo-mcp",
-        "hanzo-mcp",
-        "--allow-path",
-        "/path/to/project",
-        "--name",
-        "custom-hanzo",
-        "--transport",
-        "stdio"
-      ]
-    }
-  }
-}
-```
-
-### Configuring Claude Desktop System Prompt
-
-To get the best experience with Hanzo MCP, you need to add the provided system prompt to your Claude Desktop client. This system prompt guides Claude through a structured workflow for interacting with Hanzo platform services and managing project files.
-
-Follow these steps:
+### Quick Install
 
-1. Locate the system prompt file in this repository at `doc/system_prompt`
-2. Open your Claude Desktop client
-3. Create a new project or open an existing one
-4. Navigate to the "Project instructions" section in the Claude Desktop sidebar
-5. Copy the contents of `doc/system_prompt` and paste it into the "Project instructions" section
-6. Replace `{project_path}` with the actual absolute path to your project
+```bash
+# Install using uv
+uv pip install hanzo-mcp
 
-The system prompt provides Claude with:
+# Or using pip
+pip install hanzo-mcp
+```
 
-- A structured workflow for analyzing and modifying code
-- Best practices for project exploration and analysis
-- Guidelines for development, refactoring, and quality assurance
-- Special formatting instructions for mathematical content
+For detailed installation and configuration instructions, please refer to [INSTALL.md](./doc/INSTALL.md).
 
-This step is crucial as it enables Claude to follow a consistent approach when helping you with code modifications.
+Of course, you can also read [USEFUL_PROMPTS](./doc/USEFUL_PROMPTS.md) for some inspiration on how to use hanzo-mcp.
 
 ## Security
 
@@ -122,6 +67,48 @@ This implementation follows best practices for securing access to your filesyste
 
 ## Development
 
+### Setup Development Environment
+
+```bash
+# Clone the repository
+git clone https://github.com/hanzoai/mcp.git
+cd mcp
+
+# Install Python 3.13 using uv
+make install-python
+
+# Setup virtual environment and install dependencies
+make setup
+
+# Or install with development dependencies
+make install-dev
+```
+
+### Testing
+
+```bash
+# Run tests
+make test
+
+# Run tests with coverage
+make test-cov
+```
+
+### Building and Publishing
+
+```bash
+# Build package
+make build
+
+# Publish to PyPI
+make publish
+
+# Publish to Test PyPI
+make publish-test
+```
+
+### Contributing
+
 To contribute to this project:
 
 1. Fork the repository

hanzo_mcp-0.1.30/hanzo_mcp/__init__.py

@@ -0,0 +1,3 @@
+"""Hanzo MCP - Implementation of Hanzo capabilities using MCP."""
+
+__version__ = "0.1.8"

{hanzo_mcp-0.1.25 → hanzo_mcp-0.1.30}/hanzo_mcp/cli.py

@@ -7,13 +7,13 @@ import sys
 from pathlib import Path
 from typing import Any, cast
 
-from hanzo_mcp.server import HanzoMCPServer
+from hanzo_mcp.server import HanzoServer
 
 
 def main() -> None:
     """Run the CLI for the Hanzo MCP server."""
     parser = argparse.ArgumentParser(
-        description="MCP server for accessing Hanzo APIs and Platform capabilities"
+        description="MCP server implementing Hanzo capabilities"
     )
 
     _ = parser.add_argument(
@@ -25,20 +25,63 @@ def main() -> None:
 
     _ = parser.add_argument(
         "--name",
-        default="hanzo",
-        help="Name of the MCP server (default: hanzo)",
+        default="claude-code",
+        help="Name of the MCP server (default: claude-code)",
     )
 
     _ = parser.add_argument(
         "--allow-path",
         action="append",
         dest="allowed_paths",
-        help="Add an allowed path (can be specified multiple times)",
+        help="Add an allowed path (can be specified multiple times, defaults to user's home directory)",
     )
 
     _ = parser.add_argument(
         "--project-dir", dest="project_dir", help="Set the project directory to analyze"
     )
+    
+    _ = parser.add_argument(
+        "--agent-model",
+        dest="agent_model",
+        help="Specify the model name in LiteLLM format (e.g., 'openai/gpt-4o', 'anthropic/claude-3-sonnet')"
+    )
+    
+    _ = parser.add_argument(
+        "--agent-max-tokens",
+        dest="agent_max_tokens",
+        type=int,
+        help="Specify the maximum tokens for agent responses"
+    )
+    
+    _ = parser.add_argument(
+        "--agent-api-key",
+        dest="agent_api_key",
+        help="Specify the API key for the LLM provider (for development/testing only)"
+    )
+    
+    _ = parser.add_argument(
+        "--agent-max-iterations",
+        dest="agent_max_iterations",
+        type=int,
+        default=10,
+        help="Maximum number of iterations for agent (default: 10)"
+    )
+    
+    _ = parser.add_argument(
+        "--agent-max-tool-uses",
+        dest="agent_max_tool_uses",
+        type=int,
+        default=30,
+        help="Maximum number of total tool uses for agent (default: 30)"
+    )
+    
+    _ = parser.add_argument(
+        "--enable-agent-tool",
+        dest="enable_agent_tool",
+        action="store_true",
+        default=False,
+        help="Enable the agent tool (disabled by default)"
+    )
 
     _ = parser.add_argument(
         "--install",
@@ -53,6 +96,12 @@ def main() -> None:
     install: bool = cast(bool, args.install)
     transport: str = cast(str, args.transport)
     project_dir: str | None = cast(str | None, args.project_dir)
+    agent_model: str | None = cast(str | None, args.agent_model)
+    agent_max_tokens: int | None = cast(int | None, args.agent_max_tokens)
+    agent_api_key: str | None = cast(str | None, args.agent_api_key)
+    agent_max_iterations: int = cast(int, args.agent_max_iterations)
+    agent_max_tool_uses: int = cast(int, args.agent_max_tool_uses)
+    enable_agent_tool: bool = cast(bool, args.enable_agent_tool)
     allowed_paths: list[str] = (
         cast(list[str], args.allowed_paths) if args.allowed_paths else []
     )
@@ -61,22 +110,44 @@ def main() -> None:
         install_claude_desktop_config(name, allowed_paths)
         return
 
-    # If no allowed paths are specified, use the current directory
+    # If no allowed paths are specified, use the user's home directory
     if not allowed_paths:
-        allowed_paths = [os.getcwd()]
+        allowed_paths = [str(Path.home())]
 
     # If project directory is specified, add it to allowed paths
     if project_dir and project_dir not in allowed_paths:
         allowed_paths.append(project_dir)
 
+    # Set project directory as initial working directory if provided
+    if project_dir:
+        # Expand user paths
+        project_dir = os.path.expanduser(project_dir)
+        # Make absolute
+        if not os.path.isabs(project_dir):
+            project_dir = os.path.abspath(project_dir)
+
+    # If no specific project directory, use the first allowed path
+    elif allowed_paths:
+        project_dir = allowed_paths[0]
+
     # Run the server
-    server = HanzoMCPServer(name=name, allowed_paths=allowed_paths)
+    server = HanzoServer(
+        name=name, 
+        allowed_paths=allowed_paths,
+        project_dir=project_dir,  # Pass project_dir for initial working directory
+        agent_model=agent_model,
+        agent_max_tokens=agent_max_tokens,
+        agent_api_key=agent_api_key,
+        agent_max_iterations=agent_max_iterations,
+        agent_max_tool_uses=agent_max_tool_uses,
+        enable_agent_tool=enable_agent_tool
+    )
     # Transport will be automatically cast to Literal['stdio', 'sse'] by the server
     server.run(transport=transport)
 
 
 def install_claude_desktop_config(
-    name: str = "hanzo", allowed_paths: list[str] | None = None
+    name: str = "claude-code", allowed_paths: list[str] | None = None
 ) -> None:
     """Install the server configuration in Claude Desktop.
 

{hanzo_mcp-0.1.25 → hanzo_mcp-0.1.30}/hanzo_mcp/server.py

@@ -1,4 +1,4 @@
-"""MCP server for accessing Hanzo APIs and Platform capabilities."""
+"""MCP server implementing Hanzo capabilities."""
 
 from typing import Literal, cast, final
 
@@ -12,21 +12,35 @@ from hanzo_mcp.tools.shell.command_executor import CommandExecutor
 
 
 @final
-class HanzoMCPServer:
-    """MCP server for accessing Hanzo APIs and Platform capabilities."""
+class HanzoServer:
+    """MCP server implementing Hanzo capabilities."""
 
     def __init__(
         self,
-        name: str = "hanzo",
+        name: str = "claude-code",
         allowed_paths: list[str] | None = None,
+        project_dir: str | None = None,
         mcp_instance: FastMCP | None = None,
+        agent_model: str | None = None,
+        agent_max_tokens: int | None = None,
+        agent_api_key: str | None = None,
+        agent_max_iterations: int = 10,
+        agent_max_tool_uses: int = 30,
+        enable_agent_tool: bool = False,
     ):
         """Initialize the Hanzo server.
 
         Args:
             name: The name of the server
             allowed_paths: list of paths that the server is allowed to access
+            project_dir: Optional project directory to use as initial working directory
             mcp_instance: Optional FastMCP instance for testing
+            agent_model: Optional model name for agent tool in LiteLLM format
+            agent_max_tokens: Optional maximum tokens for agent responses
+            agent_api_key: Optional API key for the LLM provider
+            agent_max_iterations: Maximum number of iterations for agent (default: 10)
+            agent_max_tool_uses: Maximum number of total tool uses for agent (default: 30)
+            enable_agent_tool: Whether to enable the agent tool (default: False)
         """
         self.mcp = mcp_instance if mcp_instance is not None else FastMCP(name)
 
@@ -39,6 +53,11 @@ class HanzoMCPServer:
             permission_manager=self.permission_manager,
             verbose=False,  # Set to True for debugging
         )
+        
+        # If project_dir is specified, set it as initial working directory for all sessions
+        if project_dir:
+            initial_session_id = name  # Use server name as default session ID
+            self.command_executor.set_working_dir(initial_session_id, project_dir)
 
         # Initialize project analyzer
         self.project_analyzer = ProjectAnalyzer(self.command_executor)
@@ -54,13 +73,25 @@ class HanzoMCPServer:
                 self.permission_manager.add_allowed_path(path)
                 self.document_context.add_allowed_path(path)
 
+        # Store agent options
+        self.agent_model = agent_model
+        self.agent_max_tokens = agent_max_tokens
+        self.agent_api_key = agent_api_key
+        self.agent_max_iterations = agent_max_iterations
+        self.agent_max_tool_uses = agent_max_tool_uses
+        self.enable_agent_tool = enable_agent_tool
+        
         # Register all tools
         register_all_tools(
             mcp_server=self.mcp,
             document_context=self.document_context,
             permission_manager=self.permission_manager,
-            project_manager=self.project_manager,
-            project_analyzer=self.project_analyzer,
+            agent_model=self.agent_model,
+            agent_max_tokens=self.agent_max_tokens,
+            agent_api_key=self.agent_api_key,
+            agent_max_iterations=self.agent_max_iterations,
+            agent_max_tool_uses=self.agent_max_tool_uses,
+            enable_agent_tool=self.enable_agent_tool,
         )
 
     def run(self, transport: str = "stdio", allowed_paths: list[str] | None = None):
@@ -86,13 +117,13 @@ def main():
     import argparse
 
     parser = argparse.ArgumentParser(
-        description="MCP server for accessing Hanzo APIs and Platform capabilities"
+        description="MCP server implementing Hanzo capabilities"
     )
 
     _ = parser.add_argument(
         "--name",
-        default="hanzo",
-        help="Name of the MCP server (default: hanzo)",
+        default="claude-code",
+        help="Name of the MCP server (default: claude-code)",
     )
 
     _ = parser.add_argument(
@@ -117,7 +148,7 @@ def main():
     allowed_paths: list[str] | None = args.allowed_paths
 
     # Create and run the server
-    server = HanzoMCPServer(name=name, allowed_paths=allowed_paths)
+    server = HanzoServer(name=name, allowed_paths=allowed_paths)
     server.run(transport=transport, allowed_paths=allowed_paths or [])